Compare commits

..
Author SHA1 Message Date
jason.woltje 5d86b8fa93 Merge M3: release model and safe updates
Closes #10, closes #11, closes #12, closes #13
2026-09-02 20:24:33 -05:00
jason.woltje 35ea464661 docs(release): release model usage + recorded M3 drills (#13)
- README: Release model section (package/activate/rollback/status,
  pointer + append-only log, gate-then-flip guarantee)
- BUILD-LOG Phase 7: drills recorded (update, refusal, rollback),
  two harness/product corrections documented

Drill evidence: 0.0.3 -> 0.0.4 update with unchanged config checksum and
green verify; fault-injected refusal left pointer untouched; health-gated
rollback restored 0.0.3; full append-only event history.

Closes #13
2026-09-02 20:24:33 -05:00
jason.woltje e87ecdb3e5 test(release): release-layer selftests (#12)
14 cases: RELEASE validation (valid/invalid/missing), tag consistency,
status on empty state, fault-injected refusal with no pointer + single
valid refusal log line, healthy activation, pointer fields, repeat
activation append-only log, rollback-without-previous refusal.

Harness fix learned the hard way: restore RELEASE from backup inline
after the missing-file case (mv-back restored the mutated file); single
exit trap self-heals the repo state.

Closes #12
2026-09-02 20:23:18 -05:00
jason.woltje a947db7bfd feat(release): package/activate/rollback/status with health gate (#11)
- activate: image-presence pre-check + M2 task-runner health gate
  (tasks/hello-marker.json exact marker) before atomic pointer replace
  (tmp+rename); every attempt appended to activation-log.jsonl
- --fault-injection flips the health expectation to prove the refusal path
- rollback: health-gated re-activation of the previous activated imageTag
  from the log; refuses when the image is gone or no previous exists
- status: release, tag, pointer, recent log; safe on empty state
- state lives under <dataRoot>/state/ (config-independent, reset-scoped)

Verified: activate OK; fault-injected refuse with pointer unchanged;
rollback-without-previous refuse.

Closes #11
2026-09-02 20:20:32 -05:00
jason.woltje a35ea62ab1 feat(release): RELEASE identity + image tag single-sourcing (#10)
- RELEASE file: single source of release version (0.0.X until declared stable)
- common.sh load_release(): validates version, derives
  MOSAIC_IMAGE_TAG=mosaic-poc-agent:<pi>-r<release> from the pinned pi dep
- compose.yaml: image tag is required env; build/hello/verify call load_release
- verify.sh derives the image name instead of hardcoding it
- package.json version aligned to the same 0.0.X line

Closes #10
2026-09-02 20:18:17 -05:00
jason.woltje f6c93dcb5c Merge M2: mission and task abstraction
Closes #6, closes #7, closes #8, closes #9
2026-09-02 19:58:53 -05:00
jason.woltje 7f0408d417 feat(task): fixture mission/task, docs, recorded M2 E2E (#9)
- missions/hello.json + tasks/hello-marker.json committed fixtures
- README: Missions & tasks section (usage, run record layout, M2 scope note)
- BUILD-LOG: Phase 6 before/after entries

E2E: fixture run succeeded with exactly MOSAIC_HELLO_OK; wrong expectExact
recorded status failed (expect-mismatch) and exited 1; runs listed; config
checksum unchanged.

Closes #9
2026-09-02 19:58:53 -05:00
jason.woltje cfd2a19bd7 test(task): mission/task selftests — schema negatives + live runs (#8)
18 cases: validation negatives (unknown keys, versions, ids, prompt,
expectExact NUL, timeout range, missing/invalid mission, writes-nothing)
plus live cases: exact-marker success, wrong expectExact fails, distinct
run dirs, result.json contents, list output.

Closes #8
2026-09-02 19:57:34 -05:00
jason.woltje d2a9e26395 feat(task): mission/task schemas, validation, and run runner (#6) (#7)
- scripts/mosaic-task.mjs: validate | run | list
- Strict v1 schemas: unknown keys rejected; ids/prompt/expectExact/
  timeoutSeconds bounds enforced; optional mission file resolved against
  the task file and validated too
- run: executes through the config-driven container path with stdin
  detached (issue #5 class), SIGKILL timeout (default 120s), trimmed
  response capture
- Immutable run records under <dataRoot>/runs/r-<utcstamp>-<rand>/:
  task.json + mission.json snapshots (write-once), stderr.txt, result.json
- expectExact gate: mismatch -> status failed, exit 1; result.json is
  always written
- scripts/run-task.sh: load_config + bootstrap_runtime_dir before exec
- M2 scope: mission directives are snapshotted for provenance, not yet
  injected into the runtime prompt (later policy layer)

Closes #6, closes #7
2026-09-02 19:56:29 -05:00
jason.woltje 0734b1f3a5 fix(launcher): detach stdin on agent container run (#5)
pi print mode reads piped stdin until EOF; an attached terminal stdin
blocked the one-shot run forever. Automated contexts (closed stdin)
never exposed it. Request text comes from the compose command.

Proven: tail -f /dev/null | scripts/hello.sh now returns MOSAIC_HELLO_OK
in ~4s (previously timed out at 30s); verify.sh remains green.

Closes #5
2026-09-02 19:50:07 -05:00
jason.woltje ce6420f3de Merge M1: configuration-driven Hello World
Closes #1, closes #2, closes #3, closes #4
2026-09-02 18:32:50 -05:00
jason.woltje 81f58b15c8 fix(launcher): ensure configured data root before verify mount; docs for M1 (#4)
- verify.sh now calls bootstrap_runtime_dir after load_config; previously a
  reset-then-verify flow let Docker auto-create a root-owned mount source
- common.sh: fail with clear guidance when data root exists but is not writable
- README: configuration section, bootstrap usage, selftest entry point
- BUILD-LOG: Phase 5 entries with corrections

E2E (clean slate): 20/20 selftests; bootstrap idempotent; config-driven
hello/verify MOSAIC_HELLO_OK; negative marker exit 1; reset + rerun green;
config checksum unchanged across the entire flow.

Closes #4
2026-09-02 18:32:50 -05:00
jason.woltje 0c2113f710 test(config): sandboxed config-layer selftests (#3)
20 cases: bootstrap create/idempotency, missing config, malformed JSON,
unknown keys/version/backend/environment, relative and non-canonical
dataRoot, filesystem root, home dir, ancestor-of-config, control chars,
symlinked config file, env export resolution, validation-writes-nothing.

Closes #3
2026-09-02 18:30:46 -05:00
jason.woltje 900a506c1f feat(config): wire launcher scripts and compose to config.json (#2)
- common.sh: load_config() exports MOSAIC_DATA_ROOT/PROVIDER/MODEL; fails closed
- compose.yaml: dataRoot mount and provider/model are required env (:? errors)
- build/hello/verify load config before any mutation; no silent bootstrap
- reset.sh: target resolved from configured dataRoot; all safety checks kept

Verified: compose fails without launcher env; verify/reset fail on missing
config; config-driven hello+verify pass; symlink refusal with sandboxed
config (canary survived); config checksum unchanged across reset+rerun.

Closes #2
2026-09-02 18:30:08 -05:00
jason.woltje c3d29e796a feat(config): config module with idempotent bootstrap and strict v1 validation (#1)
- scripts/mosaic-config.mjs: bootstrap | validate | env operations
- Exclusive creation (O_EXCL 'wx'); existing config validated, never rewritten
- Strict schema: unknown keys rejected, configVersion===1, backend docker only
- dataRoot guards: absolute, canonical, not root/home/ancestor-of-config
- MOSAIC_CONFIG override for sandboxed tests; exit codes 0/2/3
- scripts/bootstrap.sh: explicit bootstrap entry point

Closes #1
2026-09-02 18:28:14 -05:00
jason.woltje c2365ae519 chore: baseline container POC and atomic foundation plan
- Containerized Pi hello-world proof (image mosaic-poc-agent:0.84.4, non-root)
- Four immutable contract fixtures loaded into a generated system prompt
- build/hello/verify/reset scripts with exact-match gating and reset safety
- Documented Pi discovery (v0.84.4, -p mode, --system-prompt, container auth)
- Append-only BUILD-LOG with corrections; deferred layers in LAYERS.md
- Architecture plan: docs/plans/2026-09-02_atomic-mosaic-foundation.md
2026-09-02 18:24:36 -05:00
459 changed files with 4471 additions and 52253 deletions
+20 -125
View File
@@ -1,129 +1,24 @@
# ─────────────────────────────────────────────────────────────────────────────
# Mosaic — Environment Variables Reference
# Copy this file to .env and fill in the values for your deployment.
# Lines beginning with # are comments; optional vars are commented out.
# ─────────────────────────────────────────────────────────────────────────────
# Non-secret runtime settings for the mosaic-poc-agent container.
# Copy to .env if you want to override the defaults in compose.yaml.
#
# NEVER put credentials in this file. Authentication is supplied at
# runtime only, via one of the two documented paths:
# 1. read-only mounted pi auth file (default: ~/.pi/agent/auth.json,
# override the host path with PI_AUTH_FILE)
# 2. provider API key environment variable (ZAI_API_KEY or
# ANTHROPIC_API_KEY), passed through by compose.yaml when set
# Model provider (built-in pi provider name)
PI_PROVIDER=zai
# ─── Database (PostgreSQL 17 + pgvector) ─────────────────────────────────────
# Full connection string used by the gateway, ORM, and migration runner.
# Port 5433 avoids conflict with a host-side PostgreSQL instance.
DATABASE_URL=postgresql://mosaic:mosaic@localhost:5433/mosaic
# Model ID within the provider
PI_MODEL=glm-5.3-flash
# Docker Compose host-port override for the PostgreSQL container (default: 5433)
# PG_HOST_PORT=5433
# Optional: alternative host path of the pi credential file mounted
# read-only at /home/node/.pi/agent/auth.json in the container
#PI_AUTH_FILE=/home/jwoltje/.pi/agent/auth.json
# ─── Queue (Valkey 8 / Redis-compatible) ─────────────────────────────────────
# Port 6380 avoids conflict with a host-side Redis/Valkey instance.
VALKEY_URL=redis://localhost:6380
# Docker Compose host-port override for the Valkey container (default: 6380)
# VALKEY_HOST_PORT=6380
# ─── Gateway ─────────────────────────────────────────────────────────────────
# TCP port the NestJS/Fastify gateway listens on (default: 4000)
GATEWAY_PORT=4000
# Comma-separated list of allowed CORS origins.
# Must include the web app origin in production.
GATEWAY_CORS_ORIGIN=http://localhost:3000
# ─── Auth (BetterAuth) ───────────────────────────────────────────────────────
# REQUIRED — random secret used to sign sessions and tokens.
# Generate with: openssl rand -base64 32
BETTER_AUTH_SECRET=change-me-to-a-random-32-char-string
# Public base URL of the gateway (used by BetterAuth for callback URLs)
BETTER_AUTH_URL=http://localhost:4000
# ─── Web App (Next.js) ───────────────────────────────────────────────────────
# Public gateway URL — accessible from the browser, not just the server.
NEXT_PUBLIC_GATEWAY_URL=http://localhost:4000
# ─── OpenTelemetry ───────────────────────────────────────────────────────────
# OTLP HTTP endpoint (otel-collector or any OpenTelemetry-compatible backend)
OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318
# Service name shown in traces
OTEL_SERVICE_NAME=mosaic-gateway
# ─── AI Providers ────────────────────────────────────────────────────────────
# Ollama (local models — set OLLAMA_BASE_URL to enable)
# OLLAMA_BASE_URL=http://localhost:11434
# OLLAMA_HOST is a legacy alias for OLLAMA_BASE_URL
# OLLAMA_HOST=http://localhost:11434
# Comma-separated list of Ollama model IDs to register (default: llama3.2,codellama,mistral)
# OLLAMA_MODELS=llama3.2,codellama,mistral
# OpenAI — required for embedding and log-summarization features
# OPENAI_API_KEY=sk-...
# Custom providers — JSON array of provider configs
# Format: [{"id":"<id>","baseUrl":"<url>","apiKey":"<key>","models":[{"id":"<model-id>","name":"<label>"}]}]
# MOSAIC_CUSTOM_PROVIDERS=
# ─── Embedding Service ───────────────────────────────────────────────────────
# OpenAI-compatible embeddings endpoint (default: OpenAI)
# EMBEDDING_API_URL=https://api.openai.com/v1
# EMBEDDING_MODEL=text-embedding-3-small
# ─── Log Summarization Service ───────────────────────────────────────────────
# OpenAI-compatible chat completions endpoint for log summarization (default: OpenAI)
# SUMMARIZATION_API_URL=https://api.openai.com/v1
# SUMMARIZATION_MODEL=gpt-4o-mini
# Cron schedule for summarization job (default: every 6 hours)
# SUMMARIZATION_CRON=0 */6 * * *
# Cron schedule for log tier management (default: daily at 03:00)
# TIER_MANAGEMENT_CRON=0 3 * * *
# ─── Agent ───────────────────────────────────────────────────────────────────
# Filesystem sandbox root for agent file tools (default: process.cwd())
# AGENT_FILE_SANDBOX_DIR=/var/lib/mosaic/sandbox
# Comma-separated list of tool names available to non-admin users.
# Leave unset to allow all tools for all authenticated users.
# AGENT_USER_TOOLS=read_file,list_directory,search_files
# System prompt injected into every agent session (optional)
# AGENT_SYSTEM_PROMPT=You are a helpful assistant.
# ─── MCP Servers ─────────────────────────────────────────────────────────────
# JSON array of MCP server configs — set to enable MCP tool integration.
# Each entry: {"name":"<id>","url":"<http-or-sse-url>"}
# MCP_SERVERS=[{"name":"my-mcp","url":"http://localhost:3100/sse"}]
# ─── Coordinator ─────────────────────────────────────────────────────────────
# Root directory used to scope coordinator (worktree/repo) operations.
# Defaults to the monorepo root auto-detected from process.cwd().
# MOSAIC_WORKSPACE_ROOT=/home/user/projects/mosaic
# ─── Discord Plugin (optional — set DISCORD_BOT_TOKEN to enable) ─────────────
# DISCORD_BOT_TOKEN=
# DISCORD_GUILD_ID=
# DISCORD_GATEWAY_URL=http://localhost:4000
# ─── Telegram Plugin (optional — set TELEGRAM_BOT_TOKEN to enable) ───────────
# TELEGRAM_BOT_TOKEN=
# TELEGRAM_GATEWAY_URL=http://localhost:4000
# ─── Authentik SSO (optional — set AUTHENTIK_CLIENT_ID to enable) ────────────
# AUTHENTIK_ISSUER=https://auth.example.com/application/o/mosaic/
# AUTHENTIK_CLIENT_ID=
# AUTHENTIK_CLIENT_SECRET=
# Optional: documented env-var auth alternative (secret! set in your
# shell or a gitignored .env, never commit)
#ZAI_API_KEY=
#ANTHROPIC_API_KEY=
+7 -10
View File
@@ -1,11 +1,8 @@
logs/
node_modules
dist
.turbo
.next
coverage
# build/deps
node_modules/
# runtime credentials — never commit, never copy into the image
.env
.env.local
*.tsbuildinfo
.pnpm-store
docs/reports/
secrets/
# generated runtime state lives in /home/jwoltje/.mosaic-dev (outside this project)
-1
View File
@@ -1 +0,0 @@
npx lint-staged
-1
View File
@@ -1 +0,0 @@
pnpm typecheck && pnpm lint && pnpm format:check
-9
View File
@@ -1,9 +0,0 @@
{
"**/*.{ts,tsx,js,jsx}": [
"prettier --write",
"eslint --fix"
],
"**/*.{json,md,yaml,yml}": [
"prettier --write"
]
}
-78
View File
@@ -1,78 +0,0 @@
# Repo Mosaic Linkage
This repository is attached to the machine-wide Mosaic framework.
## Load Order for Agents
1. `~/.config/mosaic/STANDARDS.md`
2. `AGENTS.md` (this repository)
3. `.mosaic/repo-hooks.sh` (repo-specific automation hooks)
## Purpose
- Keep universal standards in `~/.config/mosaic`
- Keep repo-specific behavior in this repo
- Avoid copying large runtime configs into each project
## Optional Quality Rails
Use `.mosaic/quality-rails.yml` to track whether quality rails are enabled for this repo.
Apply a template:
```bash
~/.config/mosaic/bin/mosaic-quality-apply --template <template> --target .
```
Verify enforcement:
```bash
~/.config/mosaic/bin/mosaic-quality-verify --target .
```
## Optional Matrix Orchestrator Rail
Repo-local orchestrator state lives in `.mosaic/orchestrator/`.
Run one cycle:
```bash
~/.config/mosaic/bin/mosaic-orchestrator-matrix-cycle
~/.config/mosaic/bin/mosaic-orchestrator-run --once
```
Run continuously:
```bash
~/.config/mosaic/bin/mosaic-orchestrator-run --poll-sec 10
```
Bridge events to Matrix:
```bash
~/.config/mosaic/bin/mosaic-orchestrator-matrix-publish
~/.config/mosaic/bin/mosaic-orchestrator-matrix-consume
```
Run until queue is drained (syncs from `docs/tasks.md` first):
```bash
~/.config/mosaic/bin/mosaic-orchestrator-drain
```
Set worker command if auto-detect does not match your CLI:
```bash
export MOSAIC_WORKER_EXEC="codex -p"
# or
export MOSAIC_WORKER_EXEC="opencode -p"
```
Use repo helper (foreground or detached):
```bash
bash scripts/agent/orchestrator-daemon.sh drain
bash scripts/agent/orchestrator-daemon.sh start
bash scripts/agent/orchestrator-daemon.sh status
bash scripts/agent/orchestrator-daemon.sh stop
```
-18
View File
@@ -1,18 +0,0 @@
{
"enabled": false,
"transport": "matrix",
"matrix": {
"control_room_id": "",
"workspace_id": "",
"homeserver_url": "",
"access_token": "",
"bot_user_id": ""
},
"worker": {
"runtime": "codex",
"command_template": "bash scripts/agent/orchestrator-worker.sh {task_file}",
"timeout_seconds": 7200,
"max_attempts": 1
},
"quality_gates": ["pnpm lint", "pnpm typecheck", "pnpm test"]
}
-1
View File
@@ -1 +0,0 @@
-4
View File
@@ -1,4 +0,0 @@
{
"last_published_line": 0,
"since": null
}
-14
View File
@@ -1,14 +0,0 @@
{
"schema_version": 1,
"mission_id": "mvp-20260312",
"name": "MVP",
"description": "",
"project_path": "/home/jwoltje/src/mosaic-mono-v1",
"created_at": "2026-03-13T00:44:02Z",
"status": "active",
"task_prefix": "",
"quality_gates": "",
"milestone_version": "0.0.1",
"milestones": [],
"sessions": []
}
-19
View File
@@ -1,19 +0,0 @@
{
"generated_at": "2026-03-13T00:44:51Z",
"runtime": "claude",
"mission_id": "mvp-20260312",
"mission_name": "MVP",
"project_path": "/home/jwoltje/src/mosaic-mono-v1",
"quality_gates": "",
"current_milestone": {
"id": "",
"name": ""
},
"next_task": "",
"progress": {
"tasks_done": 0,
"tasks_total": 0,
"pct": 0
},
"current_branch": ""
}
-1
View File
@@ -1 +0,0 @@
-8
View File
@@ -1,8 +0,0 @@
{
"session_id": "claude-20260312-194506-357136",
"runtime": "claude",
"pid": 357136,
"started_at": "2026-03-13T00:45:06Z",
"project_path": "/home/jwoltje/src/mosaic-mono-v1",
"milestone_id": ""
}
-4
View File
@@ -1,4 +0,0 @@
{
"running_task_id": null,
"updated_at": null
}
-3
View File
@@ -1,3 +0,0 @@
{
"tasks": []
}
-10
View File
@@ -1,10 +0,0 @@
enabled: false
template: ""
# Set enabled: true and choose one template:
# - typescript-node
# - typescript-nextjs
# - monorepo
#
# Apply manually:
# ~/.mosaic/bin/mosaic-quality-apply --template <template> --target <repo>
-17
View File
@@ -1,17 +0,0 @@
#!/usr/bin/env bash
# Optional repo-specific hooks used by scripts/agent/*.sh
# Called by session-start.sh
# mosaic_hook_session_start() {
# echo "Run repo-specific startup checks"
# }
# Called by critical.sh
# mosaic_hook_critical() {
# echo "Run repo-specific critical queries"
# }
# Called by session-end.sh
# mosaic_hook_session_end() {
# echo "Run repo-specific end-of-session checks"
# }
-1
View File
@@ -1 +0,0 @@
@mosaic:registry=https://git.mosaicstack.dev/api/packages/mosaic/npm/
-7
View File
@@ -1,7 +0,0 @@
pnpm-lock.yaml
**/next-env.d.ts
**/dist
**/node_modules
**/drizzle
**/.next
.claude/
-6
View File
@@ -1,6 +0,0 @@
{
"singleQuote": true,
"trailingComma": "all",
"semi": true,
"printWidth": 100
}
-61
View File
@@ -1,61 +0,0 @@
variables:
- &node_image 'node:22-alpine'
- &enable_pnpm 'corepack enable'
when:
- event: [push, pull_request, manual]
# Turbo remote cache (turbo.mosaicstack.dev) is configured via Woodpecker
# repository-level environment variables (TURBO_API, TURBO_TEAM, TURBO_TOKEN).
# This avoids from_secret which is blocked on pull_request events.
# If the env vars aren't set, turbo falls back to local cache only.
steps:
install:
image: *node_image
commands:
- corepack enable
- pnpm install --frozen-lockfile
typecheck:
image: *node_image
commands:
- *enable_pnpm
- pnpm typecheck
depends_on:
- install
# lint, format, and test are independent — run in parallel after typecheck
lint:
image: *node_image
commands:
- *enable_pnpm
- pnpm lint
depends_on:
- typecheck
format:
image: *node_image
commands:
- *enable_pnpm
- pnpm format:check
depends_on:
- typecheck
test:
image: *node_image
commands:
- *enable_pnpm
- pnpm test
depends_on:
- typecheck
build:
image: *node_image
commands:
- *enable_pnpm
- pnpm build
depends_on:
- lint
- format
- test
-55
View File
@@ -1,55 +0,0 @@
# Agent Guidelines — Mosaic Stack
## Required Load Order
1. `~/.config/mosaic/SOUL.md`
2. `~/.config/mosaic/STANDARDS.md`
3. `~/.config/mosaic/AGENTS.md`
4. `~/.config/mosaic/guides/E2E-DELIVERY.md`
5. `AGENTS.md` (this file)
6. Runtime-specific guide: `~/.config/mosaic/runtime/<runtime>/RUNTIME.md`
## Project Context
Mosaic Stack is a self-hosted, multi-user AI agent platform. TypeScript monorepo with NestJS gateway, Next.js web dashboard, Pi SDK agent runtime, and plugin architecture for Discord/Telegram.
## Package Map
| Package | Purpose | Key Dependencies |
| ------------------ | ------------------------------- | -------------------------------- |
| `apps/gateway` | NestJS API + WebSocket hub | Fastify, Socket.IO, Pi SDK, OTEL |
| `apps/web` | Next.js dashboard | React 19, Tailwind |
| `packages/types` | Shared TypeScript contracts | class-validator |
| `packages/db` | Drizzle ORM schema + migrations | drizzle-orm, postgres |
| `packages/auth` | BetterAuth configuration | better-auth, @mosaic/db |
| `packages/brain` | Data layer (PG-backed) | @mosaic/db |
| `packages/queue` | Valkey task queue + MCP | ioredis |
| `packages/coord` | Mission coordination | @mosaic/queue |
| `packages/cli` | Unified CLI + Pi TUI | Ink, Pi SDK |
| `plugins/discord` | Discord channel plugin | discord.js |
| `plugins/telegram` | Telegram channel plugin | Telegraf |
## Architecture Rules
1. Gateway is the single API surface — all clients connect through it
2. Pi SDK is ESM-only — gateway and CLI must use ESM
3. Socket.IO typed events defined in `@mosaic/types` enforce compile-time contracts
4. OTEL auto-instrumentation loads before NestJS bootstrap
5. BetterAuth manages auth tables; schema defined in `@mosaic/db`
6. Docker Compose provides PG (5433), Valkey (6380), OTEL Collector (4317/4318), Jaeger (16686)
7. Explicit `@Inject()` decorators required in NestJS (tsx/esbuild doesn't emit decorator metadata)
## Development Workflow
```bash
docker compose up -d # Infrastructure
pnpm install # Dependencies
pnpm typecheck && pnpm lint && pnpm format:check # Quality gates
```
## Repo-Specific Notes
- DTOs in `*.dto.ts` files at module boundaries
- ESM everywhere (`"type": "module"`, `.js` extensions in imports)
- NodeNext module resolution in all tsconfigs
- Scratchpads are mandatory for non-trivial tasks
+371
View File
@@ -0,0 +1,371 @@
# Minimal Mosaic Stack container proof of concept
## Purpose
Build the smallest isolated container that can:
- launch Pi
- load a small set of Mosaic-style contract files
- send one real request to a model
- return a known response.
This is a standalone experiment. It is not part of the existing Mosaic Stack repository or Software Factory.
## Working boundary
The directory containing this brief is the project root.
### Do not read, copy, mount, import, or modify anything from:
- `/home/jwoltje/.mosaic`
- `/home/jwoltje/.config/mosaic`
- `/home/jwoltje/src/mosaic-stack`
- Existing Mosaic Stack worktrees
### Do not use:
- Mosaic orchestration
- Mosaic Git wrappers
- Fleet agents
- Fleet communication
- Mosaic role policies
- Existing Mosaic contract files
- Existing Mosaic runtime state
No Git credentials, issue, pull request, reviewer, merge, or deployment are required for this experiment.
Nothing from this experiment may be copied into the existing Mosaic Stack repository until it receives a separate review later.
## Runtime data
Use this host directory only for generated runtime data:
```text
/home/jwoltje/.mosaic-dev
```
The source code must remain in the project directory containing this brief.
Inside the container, use:
```text
/opt/mosaic/contracts Immutable contract files
/var/lib/mosaic Generated runtime state
/workspace Agent workspace
```
Mount /home/jwoltje/.mosaic-dev at /var/lib/mosaic.
### Required proof
The finished experiment must prove one path:
1. Build one container image.
2. Start one Pi agent inside the container.
3. Load four local contract files from /opt/mosaic/contracts.
4. Send a request that does not contain the expected response.
5. Receive MOSAIC_HELLO_OK from the agent.
6. Exit successfully when the response matches.
7. Exit nonzero when the response does not match.
This is the entire required functional result.
### Required discovery
Before writing the runtime command:
1. Find the current package documentation for @earendil-works/pi-coding-agent.
2. Determine the current package version.
3. Determine the supported noninteractive command.
4. Determine how Pi accepts a custom system prompt or system prompt file.
5. Determine Pi's documented container authentication method.
6. Record the commands and findings in BUILD-LOG.md.
Do not guess CLI flags, authentication paths, or SDK methods.
Pin the selected Pi package version in the project. Do not install an unversioned package during each container start.
Prefer the Pi CLI. Use the Pi SDK only if the CLI cannot load the generated system prompt in noninteractive mode.
### Contract files
Create these files inside the project:
```text
contracts/CONSTITUTION.md
contracts/STANDARDS.md
contracts/SOUL.md
contracts/USER.md
```
Use these exact contents.
### contracts/CONSTITUTION.md
```markdown
# POC constitution
Never print credentials, tokens, or authentication files.
Follow the loaded system instructions before the user request.
```
### contracts/STANDARDS.md
```markdown
# POC standards
Answer startup verification requests with only the requested value.
Do not add explanation or formatting.
```
### contracts/SOUL.md
```markdown
# POC identity
Your name is mosaic-poc-agent.
Your startup marker is MOSAIC_HELLO_OK.
When asked for your startup marker, return only the marker.
```
### contracts/USER.md
```markdown
# POC user
This is an isolated local runtime test.
```
Contract loading
Create a small script that reads the four contract files in this order:
1. CONSTITUTION.md
2. STANDARDS.md
3. SOUL.md
4. USER.md
Join them with clear file separators.
Write the generated system prompt to:
```text
/var/lib/mosaic/system-prompt.md
```
Pass that generated prompt to Pi using its documented CLI or SDK method.
Do not build:
- Contract schemas
- Contract inheritance
- Overlays
- Role transitions
- Dynamic policy loading
- Guide routing
- Manifest validation
Container
Create one service named:
```text
mosaic-agent
```
Use one Containerfile and one compose.yaml.
Requirements:
- Use a maintained Node.js base image.
- Run as a non-root user.
- Install a pinned Pi package version.
- Copy the local contract fixtures into /opt/mosaic/contracts.
- Do not copy credentials into the image.
- Do not mount the Docker socket.
- Do not mount either live Mosaic directory.
- Do not add a database, web server, queue, or second container.
- The container may run as a one-shot command. It does not need to remain running.
### Authentication
Use Pi's documented authentication mechanism.
Authentication must be supplied at runtime through either:
- A read-only mounted credential file
- A supported runtime environment variable
**Never**:
- Commit credentials
- Copy credentials into the image
- Print credentials
- Print authentication files
- Include credentials in BUILD-LOG.md
- Store credentials under the project directory
Provide .env.example only for non-secret settings such as model or provider names.
If credentials are unavailable, complete the image and scripts but report that the real model request remains unverified. Do not fake the response.
### Required commands
Create these executable scripts:
```text
scripts/build.sh
scripts/hello.sh
scripts/verify.sh
scripts/reset.sh
```
### scripts/build.sh
Build the container image using Docker Compose.
### scripts/hello.sh
Run the mosaic-agent service as a one-shot container.
Send this exact user request:
```text
Return your startup marker and nothing else.
```
The request must not contain MOSAIC_HELLO_OK.
Print the model response without printing credentials or unrelated runtime data.
### scripts/verify.sh
Run the complete test.
**It must**:
1. Build or confirm the image is built.
2. Run the agent request.
3. Remove surrounding whitespace from the response.
4. Compare the response with MOSAIC_HELLO_OK.
5. Exit 0 only when they match exactly.
6. Exit nonzero with a clear error when they do not match.
### scripts/reset.sh
Delete generated POC state only when all checks pass:
1. The resolved path is exactly /home/jwoltje/.mosaic-dev.
2. The path is not a symbolic link.
3. The directory contains a .mosaic-poc-root ownership marker created by this project.
Refuse to delete anything if a check fails.
## Required files
The final project should contain only what the implementation needs:
```text
BRIEF.md
BUILD-LOG.md
README.md
LAYERS.md
Containerfile
compose.yaml
package.json
package-lock.json
.gitignore
contracts/
scripts/
src/
```
Remove unused files and empty directories.
Build log
Create BUILD-LOG.md.
Treat it as append-only.
Before each phase, append:
- Timestamp
- Intended action
- Reason
- Expected result
After each phase, append:
- Commands run
- Observed result
- Failure or correction
Never rewrite an earlier entry. Add a correction as a new entry.
Do not record credentials.
Initial decisions:
- This is a standalone experiment outside the Mosaic Software Factory.
- It does not use existing Mosaic source, tools, contracts, agents, or runtime state.
- The first proof uses one Pi agent and four small local contract files.
- The only required model result is MOSAIC_HELLO_OK.
- Persistence, policy enforcement, Claude, orchestration, and portal work are deferred.
## Acceptance criteria
The experiment passes when:
1. scripts/build.sh exits 0.
2. The image contains the four local contract files.
3. The image contains no credentials.
4. The container has no mounts from ~/.mosaic or ~/.config/mosaic.
5. scripts/hello.sh performs a real model request.
6. The request does not contain the expected marker.
7. The agent returns exactly MOSAIC_HELLO_OK.
8. scripts/verify.sh exits 0.
9. Changing the expected value makes scripts/verify.sh exit nonzero.
10. scripts/reset.sh refuses unsafe paths.
11. Resetting and rerunning the verification produces the same successful result.
## Deferred layers
Document these in LAYERS.md. Do not implement them.
- L0: Container builds and returns MOSAIC_HELLO_OK.
- L1: Persist and resume a named Pi session.
- L2: Add a fixed tool permission policy.
- L3: Load full versioned contract bundles.
- L4: Add Claude as a second runtime.
- L5: Add multiple agents and communication.
- L6: Add orchestration, knowledge storage, and portal features.
## Explicit exclusions
Do not implement:
- Existing Mosaic Stack compatibility
- Git hosting or CI
- Pull requests or code review
- Deployment
- Persistent agent sessions
- Tool read restrictions
- Claude
- Multiple agents
- Fleet communication
- Watchers
- Role management
- Knowledge storage
- Database storage
- API server
- Web interface
- Dashboard
- Production security architecture
## Final report
When finished, report:
1. Files created.
2. Pi package version.
3. Exact build command.
4. Exact verification command.
5. Verification output with credentials removed.
6. Whether the real model request passed.
7. Any remaining failure.
8. Anything implemented beyond this brief.
Do not describe the experiment as production-ready.
+195
View File
@@ -0,0 +1,195 @@
# BUILD-LOG
Append-only build log for the Minimal Mosaic Stack container proof of concept.
Each phase records the plan before it runs and the observed result after it runs.
No credentials are recorded in this file.
---
## Phase 1: Pi package discovery
### Entry 1.1 — before
- Timestamp: 2026-02-02 (session start, local)
- Intended action: Locate the current package documentation for `@earendil-works/pi-coding-agent`, determine the current version, the supported noninteractive command, the custom system prompt mechanism, and the documented container authentication method.
- Reason: The brief forbids guessing CLI flags, authentication paths, or SDK methods; all runtime commands must be derived from the package documentation.
- Expected result: Documented answers for all five discovery questions, recorded below, with the Pi package version pinned in the project.
### Entry 1.2 — after
- Timestamp: 2026-02-02
- Commands run:
- Read `README.md` of the locally installed `@earendil-works/pi-coding-agent` package
- Read `docs/containerization.md`, excerpts of `docs/models.md` and `docs/providers.md`
- `grep '"version"' .../pi-coding-agent/package.json`
- `npm view @earendil-works/pi-coding-agent version`
- Inspected host auth store structure (keys only, values never printed)
- Observed result:
1. **Package documentation**: full docs ship inside the installed package (`README.md` plus a `docs/` directory including `docs/containerization.md`).
2. **Current package version**: `0.84.4` — the locally installed version and the npm registry latest are identical. Selected version to pin: `0.84.4`.
3. **Supported noninteractive command**: `pi -p` / `pi --print` — "Print response and exit". Documented in the CLI Reference. Print mode also merges piped stdin into the initial prompt (not used here).
4. **Custom system prompt**: documented CLI flag `--system-prompt <text>` — "Replace default prompt (context files and skills still appended)". Because the generated contract prompt must fully control behavior, the run also passes the documented discovery-off switches (`--no-context-files`, `--no-skills`, `--no-extensions`, `--no-prompt-templates`, `--no-themes`) so nothing is appended from ambient files. (`APPEND_SYSTEM.md` / `.pi/SYSTEM.md` file conventions exist but the CLI flag is the documented noninteractive path.)
5. **Documented container authentication**: `docs/containerization.md` documents the "Plain Docker" pattern: run the whole `pi` process in the container and pass provider API keys at runtime via environment variables (example: `docker run -e ANTHROPIC_API_KEY ...`). The documented alternative credential store is `auth.json` written by `/login` under the pi agent directory (`~/.pi/agent/auth.json`).
- Decision: supply the credential at runtime as a **read-only mounted credential file** (host `auth.json`, which contains a static API-key entry for the built-in `zai` provider) mounted at the container pi agent directory, and additionally allow the documented environment-variable path (`ZAI_API_KEY` / `ANTHROPIC_API_KEY`) as an alternative. Provider and model are non-secret settings supplied via `.env` (`PI_PROVIDER=zai`, `PI_MODEL=glm-5.3-flash`).
- Failure or correction: none. Host check confirmed no API-key environment variables are exported on the host, so the read-only auth.json mount is the working runtime credential path for this experiment.
---
## Phase 2: Project scaffold
### Entry 2.1 — before
- Timestamp: 2026-02-02
- Intended action: Create the contract fixtures (exact brief contents), the contract loader (`src/load-contracts.sh`), the one-shot agent runner (`src/run-agent.sh`), the four required scripts (`scripts/build.sh`, `hello.sh`, `verify.sh`, `reset.sh`), `Containerfile`, `compose.yaml`, pinned `package.json` + `package-lock.json`, `.gitignore`, `README.md`, `LAYERS.md`.
- Reason: Implement exactly the file set the brief requires, with no extra machinery (no schemas, overlays, manifests, or policy loading).
- Expected result: A complete project whose only remaining unknown is whether the pinned image builds and the real model request returns `MOSAIC_HELLO_OK`.
### Entry 2.2 — after
- Timestamp: 2026-02-02
- Commands run: file creation; `npm install --package-lock-only --ignore-scripts` to generate the lockfile from the pinned dependency.
- Observed result: All files created; `package-lock.json` pins `@earendil-works/[email protected]` (exact, no range).
- Failure or correction: none.
---
## Phase 3: Container image build
### Entry 3.1 — before
- Timestamp: 2026-02-02
- Intended action: Run `scripts/build.sh` (Docker Compose build) to produce image `mosaic-poc-agent:0.84.4` from `node:24-bookworm-slim` with the pinned Pi, the four contract fixtures at `/opt/mosaic/contracts`, and a non-root user (uid/gid 1000).
- Reason: Phase 1 of the required proof path; `node:24-bookworm-slim` is the maintained base image used in Pi's own documented containerization example.
- Expected result: `docker compose build` exits 0 and the image contains the contracts, the runner scripts, and the pinned `pi` binary, with no credentials baked in.
### Entry 3.2 — after
- Timestamp: 2026-02-02
- Commands run: `scripts/build.sh`; `docker run --rm mosaic-poc-agent:0.84.4 --version`; `id` via `--entrypoint`; contract listing; credential file scan.
- Observed result:
- Build exit 0; image tagged `mosaic-poc-agent:0.84.4`.
- `pi --version` inside the image reports `0.84.4` (and this run also executed the contract loader successfully, writing `/var/lib/mosaic/system-prompt.md`).
- Container user is `uid=1000(node) gid=1000(node)` — non-root.
- All four contract files present at `/opt/mosaic/contracts` with read-only permissions (0555).
- Credential scan: no `auth.json` or other auth files exist in the image; `/home/node/.pi/agent/` is empty in the image.
- Failure or correction:
1. First build failed: Docker Compose expects `Dockerfile` by default; fixed by setting `build.dockerfile: Containerfile` in `compose.yaml`.
2. Second build failed: `useradd` exit 4 (uid 1000 already exists) because the maintained node image ships a `node` user at uid/gid 1000. Fixed by reusing the built-in `node` user (same 1000:1000 host mapping) instead of creating a duplicate `mosaic` user; container paths updated from `/home/mosaic/...` to `/home/node/...` in `Containerfile`, `compose.yaml`, `README.md`, `.env.example`.
---
## Phase 4: Runtime verification
### Entry 4.1 — before
- Timestamp: 2026-02-02
- Intended action: Run `scripts/hello.sh` (one-shot request: "Return your startup marker and nothing else."), then `scripts/verify.sh` (exact-match gate against `MOSAIC_HELLO_OK`), then the negative test (`EXPECTED_MARKER=MOSAIC_NOT_OK scripts/verify.sh` must exit nonzero), then the `scripts/reset.sh` safety tests and a final rerun after reset.
- Reason: Phases 27 of the required proof path plus acceptance criteria 511.
- Expected result: hello prints only the marker; verify exits 0; negative test exits nonzero; reset refuses unsafe paths and succeeds on the real path; rerun after reset reproduces the success.
### Entry 4.2 — after
- Timestamp: 2026-02-02
- Commands run: `scripts/hello.sh`; `scripts/verify.sh`; `EXPECTED_MARKER=MOSAIC_NOT_OK scripts/verify.sh`; `scripts/reset.sh` (refusal tests: missing marker, symlink with canary file, then real reset, then missing dir); `scripts/build.sh && scripts/verify.sh` after reset; `docker compose config` mount inspection.
- Observed result:
- `hello.sh`: stdout exactly `MOSAIC_HELLO_OK` — a real model request (provider `zai`, model `glm-5.3-flash`, auth via the read-only mounted auth.json credential file). The request string contains no marker.
- `verify.sh`: `PASS: response matches expected marker`, exit 0.
- Negative test: `FAIL: response does not match expected marker` (expected `MOSAIC_NOT_OK`, actual `MOSAIC_HELLO_OK`), exit 1.
- `reset.sh` refusal tests: missing marker → exit 1, nothing deleted; symlink (with canary file at the target) → exit 1, canary survived; real path with marker → removed, exit 0; missing dir → "nothing to remove", exit 0.
- Rerun after reset: build + verify → PASS, exit 0 (criterion 11).
- Resolved compose mounts: only `/home/jwoltje/.mosaic-dev → /var/lib/mosaic` (rw) and `~/.pi/agent/auth.json → /home/node/.pi/agent/auth.json` (read-only). No `~/.mosaic` or `~/.config/mosaic` mounts, no Docker socket.
- Failure or correction:
1. First hello run: the contract loader's status line was printed on stdout, mixing runtime data into the model response stream and contaminating the exact-match capture. Fixed by sending the loader's status message to stderr (`src/load-contracts.sh`), rebuilt the image, reran: stdout is exactly the model response.
- Credential check: no credential material appears in this log, in hello/verify output, or in the image (image scan found no auth files).
## Result
All 11 acceptance criteria demonstrated. The real model request passed.
---
## Phase 5: Configuration-driven Hello World (M1)
### Entry 5.1 — before
- Timestamp: 2026-09-03
- Intended action: Make the container POC configuration-driven. Baseline committed and tagged `poc-container-hello-v0`. Milestone M1 tracked in Gitea (issues #1-#4): (T1) config module with idempotent bootstrap and strict v1 validation; (T2) wire scripts and compose to config.json with fail-closed behavior; (T3) sandboxed config selftests; (T4) E2E verification and documentation.
- Reason: Per docs/plans/2026-09-02_atomic-mosaic-foundation.md — config.json must be the sole discovery entry point; updates and runs must never corrupt or invent configuration.
- Expected result: All M1 acceptance criteria pass; Hello World reproducible from configuration alone.
### Entry 5.2 — after
- Timestamp: 2026-09-03
- Commands run: `scripts/test-config.sh` (20 cases); fail-closed checks (compose without launcher env, verify/reset with missing config); `scripts/bootstrap.sh`; config-driven `scripts/hello.sh`, `scripts/verify.sh`, negative marker test, sandboxed reset symlink refusal (canary survived), real reset + bootstrap + build + verify; config checksum comparison across the entire flow.
- Observed result:
- Config selftests: 20 passed, 0 failed.
- Fail-closed confirmed: compose exits 1 without launcher env; verify/reset exit 1 on missing config before any mutation.
- Bootstrap created `~/.config/mosaic-dev/config.json` exclusively; second run validated without rewriting (content + mtime unchanged).
- Config-driven hello/verify returned exactly `MOSAIC_HELLO_OK`; verify exit 0; negative marker test exit 1.
- Reset refused symlinked dataRoot; canary file survived; real reset removed only the configured data root.
- config.json checksum unchanged across hello/verify/reset/bootstrap/build/verify.
- Failure or correction:
1. Selftest harness bug: `cfg` helper invoked without a body for the symlink case (`$2: unbound variable`). Fixed in the harness; product code unaffected.
2. E2E rerun-after-reset failure: `verify.sh` did not ensure the configured data root existed before the container mount. With the data root absent, Docker auto-created the host path as root:root, and the container's uid-1000 user could not write the generated system prompt. Fixed by calling `bootstrap_runtime_dir` in `verify.sh`; also hardened it to fail with a clear message when the data root exists but is not writable (root-owned leftover). Clean-slate E2E rerun: all steps green.
- Credential check: no credential material in config, scripts, logs, or test output.
## Result (M1)
Configuration-driven Hello World verified. `main` merged with M1 and tagged `config-hello-v1`.
---
## Phase 6: Mission and task abstraction (M2)
### Entry 6.1 — before
- Timestamp: 2026-09-03
- Intended action: Add the first mission/task layer, host-side only (Gitea milestone M2, issues #6-#9): strict v1 schemas for missions and tasks, a task runner executing through the proven config-driven container path, immutable write-once run records under `<dataRoot>/runs/`, `expectExact` gating, timeouts, selftests, fixtures, and docs.
- Reason: The foundation plan's following layer — mission (objective + directives), task (bounded unit), run (one attempt), result (immutable evidence) — must exist as data and records before any policy or multi-agent work.
- Expected result: `scripts/run-task.sh tasks/hello-marker.json` succeeds with exactly `MOSAIC_HELLO_OK`; wrong expectations fail; every run leaves an immutable record; configuration remains untouched.
### Entry 6.2 — after
- Timestamp: 2026-09-03
- Commands run: `scripts/test-task.sh` (18 cases incl. live runs); `scripts/run-task.sh validate` / `run` on committed fixtures; `node scripts/mosaic-task.mjs list`; config checksum comparison across runs.
- Observed result:
- Selftests: 18 passed, 0 failed (schema negatives; live exact-marker success; wrong expectExact fails; distinct run dirs; result.json contents; list).
- Fixture run: status `succeeded`, response exactly `MOSAIC_HELLO_OK`, mission snapshot recorded.
- Run records written once under `<dataRoot>/runs/r-<utcstamp>-<rand>/`; reruns never clobber.
- Failure or correction: none this phase.
- Credential check: no credential material in task data, run records, or logs.
## Result (M2)
Mission/task layer verified end-to-end. `main` merged with M2 and tagged `mission-task-v1`.
---
## Phase 7: Release model and safe updates (M3)
### Entry 7.1 — before
- Timestamp: 2026-09-03
- Intended action: Add the release substrate (Gitea milestone M3, issues #10-#13): RELEASE file single-sources the version (0.0.X line per owner direction), image tags derive from it, scripts/release.sh provides package/activate/rollback/status, activation is health-gated by the M2 task runner, pointer + append-only log under <dataRoot>/state/.
- Reason: The owner's top invariant — updates must never corrupt a working installation — needs a mechanism, not a convention: gate-then-flip with recorded history and rollback.
- Expected result: Update, refusal, and rollback drills all green with config checksums unchanged.
### Entry 7.2 — after
- Timestamp: 2026-09-03
- Commands run: scripts/test-release.sh (14 cases); recorded drills: update (0.0.3 -> 0.0.4 package+activate+verify), fault-injected refusal, rollback to 0.0.3.
- Observed result:
- Selftests: 14 passed, 0 failed.
- Update drill: packaged and activated r0.0.4 after exact-marker health gate; verify green under the new tag; config checksum unchanged.
- Refusal drill: health-gate fault injection -> activation refused (exit 1), pointer untouched, refusal appended to the log.
- Rollback drill: health-gated rollback to r0.0.3; pointer restored; log records package/activate/refused/rollback history append-only.
- Failure or correction:
1. release.sh initially failed with missing state/ directory (no mkdir before pointer/log writes); fixed.
2. Selftest harness mutated the repo RELEASE and restored the mutated copy (mv-back bug) plus a second trap replacing the first; fixed with inline backup restore and one self-healing exit trap. Product code unaffected.
- Credential check: no credential material in release state, logs, or drills.
## Result (M3)
Release model and safe updates verified by drills. `main` merged with M3 and tagged `release-model-v1`.
-45
View File
@@ -1,45 +0,0 @@
# CLAUDE.md — Mosaic Stack
## Project
Self-hosted, multi-user AI agent platform. TypeScript monorepo.
## Stack
- **API**: NestJS + Fastify adapter (`apps/gateway`)
- **Web**: Next.js 16 + React 19 (`apps/web`)
- **ORM**: Drizzle ORM + PostgreSQL 17 + pgvector (`packages/db`)
- **Auth**: BetterAuth (`packages/auth`)
- **Agent**: Pi SDK (`packages/agent`, `packages/cli`)
- **Queue**: Valkey 8 (`packages/queue`)
- **Build**: pnpm workspaces + Turborepo
- **CI**: Woodpecker CI
- **Observability**: OpenTelemetry → Jaeger
## Commands
```bash
pnpm typecheck # TypeScript check (all packages)
pnpm lint # ESLint (all packages)
pnpm format:check # Prettier check
pnpm test # Vitest (all packages)
pnpm build # Build all packages
# Database
pnpm --filter @mosaic/db db:push # Push schema to PG (dev)
pnpm --filter @mosaic/db db:generate # Generate migrations
pnpm --filter @mosaic/db db:migrate # Run migrations
# Dev
docker compose up -d # Start PG, Valkey, OTEL, Jaeger
pnpm --filter @mosaic/gateway exec tsx src/main.ts # Start gateway
```
## Conventions
- ESM everywhere (`"type": "module"`, `.js` extensions in imports)
- NodeNext module resolution
- Explicit `@Inject()` decorators in NestJS (tsx/esbuild doesn't support emitDecoratorMetadata)
- DTOs in `*.dto.ts` files at module boundaries
- OTEL tracing imported before NestJS bootstrap (`import './tracing.js'`)
- All three gates must pass before push: typecheck, lint, format:check
+40
View File
@@ -0,0 +1,40 @@
# Minimal Mosaic Stack POC agent image.
# Base: maintained Node.js image (same family as Pi's documented
# containerization example in docs/containerization.md).
FROM node:24-bookworm-slim
# Tools Pi's documented container image expects (bash, CA certs, git, ripgrep).
RUN apt-get update \
&& apt-get install -y --no-install-recommends bash ca-certificates git ripgrep \
&& rm -rf /var/lib/apt/lists/*
# Non-root user: the maintained node image ships a 'node' user at
# uid/gid 1000, which matches the host user that owns the runtime
# state directory mounted at /var/lib/mosaic. It is reused as-is.
# Pinned Pi install: package.json pins the exact version and
# package-lock.json is installed with npm ci. No unversioned installs.
WORKDIR /opt/app
COPY package.json package-lock.json ./
RUN npm ci --ignore-scripts
# Immutable contract fixtures (required location) and runtime scripts.
COPY contracts /opt/mosaic/contracts
COPY src /opt/mosaic/src
RUN chmod 0555 /opt/mosaic/contracts /opt/mosaic/contracts/* \
&& chmod 0555 /opt/mosaic/src /opt/mosaic/src/*.sh
# Writable state, workspace, and pi agent directory (auth.json is
# bind-mounted read-only at runtime; nothing is copied into the image).
RUN mkdir -p /var/lib/mosaic /workspace /home/node/.pi/agent \
&& chown -R node:node /var/lib/mosaic /workspace /home/node /opt/app
USER node
WORKDIR /workspace
ENV HOME=/home/node \
PATH="/opt/app/node_modules/.bin:${PATH}" \
PI_OFFLINE=1
# One-shot agent: args form the user request (default is the startup
# verification request defined in compose.yaml).
ENTRYPOINT ["/opt/mosaic/src/run-agent.sh"]
+50
View File
@@ -0,0 +1,50 @@
# LAYERS
Deferred capability layers for the Mosaic experiment. Only L0 is implemented by
this proof of concept; everything below it is documented here and deliberately
not implemented (see BRIEF.md, "Explicit exclusions").
## L0 — Implemented: container returns MOSAIC_HELLO_OK
One image (`mosaic-poc-agent:0.84.4`, built on `node:24-bookworm-slim`, non-root,
pinned Pi) runs one Pi agent one-shot. Four immutable local contract files are
loaded in fixed order into the generated system prompt
(`/var/lib/mosaic/system-prompt.md`). One real model request is sent
noninteractively; the response must equal `MOSAIC_HELLO_OK` exactly or the
verification exits nonzero. Authentication is supplied at runtime only
(read-only mounted pi auth file, or a provider API key environment variable).
## L1 — Deferred: persist and resume a named Pi session
Keep a named Pi session across container runs (`--name`, session storage under
`/var/lib/mosaic`), resume it with the documented session flags, and verify
state survives a container restart.
## L2 — Deferred: fixed tool permission policy
Add a fixed allow/deny policy for Pi tools (e.g. restricting built-in tools via
documented `--tools` / `--exclude-tools` or an extension-based permission gate),
so contract files can constrain what the agent may do, not just what it says.
## L3 — Deferred: load full versioned contract bundles
Replace the four static fixtures with versioned contract bundles: bundle
manifests, contract versions, and deterministic ordering/hashing, loaded from
an immutable bundle artifact instead of files copied at image build time.
## L4 — Deferred: Claude as a second runtime
Add a second runtime (Claude) alongside the Pi agent in the same container
stack, behind the same contract-loading path, to compare behavior across
runtimes.
## L5 — Deferred: multiple agents and communication
Run several named agents with defined roles and a communication channel between
them (message passing or shared state under `/var/lib/mosaic`).
## L6 — Deferred: orchestration, knowledge storage, and portal features
Fleet-level orchestration, knowledge storage, monitoring, and portal UI on top
of L1-L5. This is where the existing Mosaic Stack concepts would be re-evaluated
from first principles.
+173
View File
@@ -0,0 +1,173 @@
# Minimal Mosaic Stack container POC
Standalone experiment, not part of the Mosaic Stack repository or Software Factory.
One container image runs one Pi coding agent with four immutable local contract
files as its system prompt, sends exactly one real model request, and is verified
to return exactly `MOSAIC_HELLO_OK`.
## Layout
```text
BRIEF.md requirements for the original container proof
BUILD-LOG.md append-only build/verification log
LAYERS.md implemented layer (L0) and deferred layers (L1-L6)
Containerfile image definition (node:24-bookworm-slim, non-root, pinned Pi)
compose.yaml one service: mosaic-agent (one-shot; configured via env)
package.json pins @earendil-works/pi-coding-agent at exactly 0.84.4
package-lock.json resolved lockfile used by npm ci in the image
.env.example non-secret settings only (credential-file path, env-var auth)
contracts/ CONSTITUTION.md, STANDARDS.md, SOUL.md, USER.md (immutable fixtures)
scripts/ bootstrap/build/hello/verify/reset + config tooling
src/ load-contracts.sh, run-agent.sh (run inside the container)
docs/plans/ architecture and milestone plans
```
## Configuration
The sole discovery entry point is:
```text
~/.config/mosaic-dev/config.json
```
Created only by the explicit, idempotent bootstrap:
```bash
scripts/bootstrap.sh # create-if-absent; validates existing config, never rewrites
```
Minimal shape (`configVersion` 1):
```json
{
"configVersion": 1,
"environment": "development",
"dataRoot": "/home/jwoltje/.mosaic-dev",
"execution": {
"backend": "docker",
"provider": "zai",
"model": "glm-5.3-flash"
}
}
```
Rules enforced by `scripts/mosaic-config.mjs`:
- Unknown keys, unsupported versions/backends, and malformed JSON exit nonzero; nothing is modified.
- `dataRoot` must be absolute, canonical, and must not be or contain the home or configuration directory.
- Validation failures never touch config, state, or images.
- `scripts/test-config.sh` runs the sandboxed config selftests (no Docker required).
Run paths (`build/hello/verify/reset`) fail closed when configuration is missing or invalid; they never invent it.
## Missions & tasks (M2)
Missions and tasks are validated JSON data (strict schemas, version-pinned). The M2 layer is host-side only: mission directives are recorded for provenance but do not yet reach the runtime system prompt (capability/policy layer comes later).
```text
missions/hello.json objective + directives (missionVersion 1)
tasks/hello-marker.json prompt + optional mission ref + expectExact + timeout
<dataRoot>/runs/r-<id>/ immutable run record: task.json, mission.json,
stderr.txt, result.json (all write-once)
```
Usage:
```bash
scripts/run-task.sh validate tasks/hello-marker.json # strict validation, writes nothing
scripts/run-task.sh run tasks/hello-marker.json # execute; result recorded under dataRoot/runs
scripts/mosaic-task.mjs list # list runs and statuses
scripts/test-task.sh # selftests (schema negatives + live runs)
```
A run exits 0 only when its expectation is met (`expectExact` match); mismatches, nonzero agent exits, and timeouts record `status: failed` in `result.json` and exit 1. Each run gets a unique directory — rerunning never rewrites history.
## Release model (M3)
`RELEASE` single-sources the release version (0.0.X until declared stable); the image tag derives from it plus the pinned Pi version. Activation is health-gated and every event is recorded:
```bash
scripts/release.sh package # build + tag the release image
scripts/release.sh activate # health check (exact marker) -> atomic pointer swap
scripts/release.sh activate --fault-injection # prove the refusal path (drills only)
scripts/release.sh rollback # health-gated return to the previous release
scripts/release.sh status # release, tag, active pointer, recent log
scripts/test-release.sh # release selftests
```
- `<dataRoot>/state/active.json` — the activation pointer (atomic tmp+rename replace)
- `<dataRoot>/state/activation-log.jsonl` — append-only history: package / activate / refused / rollback
A failed health check never activates; the previously active release remains deployed. Updating the software therefore cannot corrupt the running installation: package beside, gate, then flip. Verified by the update/refusal/rollback drills in BUILD-LOG Phase 7.
See `docs/plans/2026-09-02_atomic-mosaic-foundation.md` for the full plan.
Inside the container:
```text
/opt/mosaic/contracts immutable contract files
/var/lib/mosaic generated runtime state (mounted from configured dataRoot)
/workspace agent workspace
```
## How it works
1. `scripts/build.sh` builds the release image (`mosaic-poc-agent:<pi>-r<release>`,
tag derived from `RELEASE` + the pinned Pi version) with Docker Compose.
2. On each run, `/opt/mosaic/src/load-contracts.sh` reads the four contract files
in fixed order (CONSTITUTION, STANDARDS, SOUL, USER), joins them with clear
separators, and writes `/var/lib/mosaic/system-prompt.md`.
3. `/opt/mosaic/src/run-agent.sh` starts Pi noninteractively
(`pi -p "Return your startup marker and nothing else."`) with
`--system-prompt "$(cat /var/lib/mosaic/system-prompt.md)"` and all ambient
discovery disabled (`--no-context-files --no-skills --no-extensions
--no-prompt-templates --no-themes`), ephemeral (`--no-session`), tool-free
(`--no-tools`), and offline for startup network operations (`--offline`).
4. `scripts/verify.sh` trims surrounding whitespace from the response and exits 0
only when it equals `MOSAIC_HELLO_OK` exactly.
## Usage
```bash
scripts/bootstrap.sh # create config.json if absent (idempotent)
scripts/build.sh # build the image
scripts/hello.sh # one-shot request; prints the model response
scripts/verify.sh # full gated test; exit 0 only on exact MOSAIC_HELLO_OK
scripts/run-task.sh # run a mission/task file (see Missions & tasks)
scripts/release.sh # package / activate / rollback / status (see Release model)
scripts/test-config.sh # fast config-layer selftests (no Docker)
scripts/test-task.sh # mission/task selftests (schema + live runs)
scripts/test-release.sh # release selftests
scripts/reset.sh # delete the configured data root (safety-checked)
```
Prove the failure path (acceptance criterion 9):
```bash
EXPECTED_MARKER=MOSAIC_NOT_OK scripts/verify.sh # must exit nonzero
```
## Authentication
Pi's documented container authentication (see the package's
`docs/containerization.md`) is used, in this order:
1. **Read-only mounted credential file** (default): the host pi auth file
`~/.pi/agent/auth.json` is bind-mounted read-only to
`/home/node/.pi/agent/auth.json`. The host file holds a static API-key
entry for the built-in `zai` provider, so no token refresh writes are needed.
2. **Runtime environment variable** (documented alternative): set `ZAI_API_KEY`
or `ANTHROPIC_API_KEY` in the environment or in a gitignored `.env`; compose
passes them through. Pi's documented precedence applies.
Credentials are never committed, never copied into the image, and never printed.
`.env.example` contains non-secret settings only.
## Boundaries honored
- No mounts of `~/.mosaic` or `~/.config/mosaic`; no Docker socket mount.
- Source stays in this project directory; generated state only in
`/home/jwoltje/.mosaic-dev` (host) and `/var/lib/mosaic` (container).
- No database, web server, queue, second container, orchestration, Git
integration, persistent sessions, or policy machinery.
+1
View File
@@ -0,0 +1 @@
0.0.3
-63
View File
@@ -1,63 +0,0 @@
{
"name": "@mosaic/gateway",
"version": "0.0.0",
"private": true,
"type": "module",
"main": "dist/main.js",
"scripts": {
"build": "tsc",
"dev": "tsx watch src/main.ts",
"lint": "eslint src",
"typecheck": "tsc --noEmit -p tsconfig.typecheck.json",
"test": "vitest run --passWithNoTests"
},
"dependencies": {
"@fastify/helmet": "^13.0.2",
"@mariozechner/pi-ai": "~0.57.1",
"@mariozechner/pi-coding-agent": "~0.57.1",
"@modelcontextprotocol/sdk": "^1.27.1",
"@mosaic/auth": "workspace:^",
"@mosaic/queue": "workspace:^",
"@mosaic/brain": "workspace:^",
"@mosaic/coord": "workspace:^",
"@mosaic/db": "workspace:^",
"@mosaic/discord-plugin": "workspace:^",
"@mosaic/log": "workspace:^",
"@mosaic/memory": "workspace:^",
"@mosaic/telegram-plugin": "workspace:^",
"@mosaic/types": "workspace:^",
"@nestjs/common": "^11.0.0",
"@nestjs/core": "^11.0.0",
"@nestjs/platform-fastify": "^11.0.0",
"@nestjs/platform-socket.io": "^11.0.0",
"@nestjs/throttler": "^6.5.0",
"@nestjs/websockets": "^11.0.0",
"@opentelemetry/auto-instrumentations-node": "^0.71.0",
"@opentelemetry/exporter-metrics-otlp-http": "^0.213.0",
"@opentelemetry/exporter-trace-otlp-http": "^0.213.0",
"@opentelemetry/resources": "^2.6.0",
"@opentelemetry/sdk-metrics": "^2.6.0",
"@opentelemetry/sdk-node": "^0.213.0",
"@opentelemetry/semantic-conventions": "^1.40.0",
"@sinclair/typebox": "^0.34.48",
"better-auth": "^1.5.5",
"class-transformer": "^0.5.1",
"class-validator": "^0.15.1",
"dotenv": "^17.3.1",
"fastify": "^5.0.0",
"node-cron": "^4.2.1",
"reflect-metadata": "^0.2.0",
"rxjs": "^7.8.0",
"socket.io": "^4.8.0",
"uuid": "^11.0.0",
"zod": "^4.3.6"
},
"devDependencies": {
"@types/node": "^22.0.0",
"@types/node-cron": "^3.0.11",
"@types/uuid": "^10.0.0",
"tsx": "^4.0.0",
"typescript": "^5.8.0",
"vitest": "^2.0.0"
}
}
@@ -1,148 +0,0 @@
import { ForbiddenException, NotFoundException } from '@nestjs/common';
import { describe, expect, it, vi } from 'vitest';
import { ConversationsController } from '../conversations/conversations.controller.js';
import { MissionsController } from '../missions/missions.controller.js';
import { ProjectsController } from '../projects/projects.controller.js';
import { TasksController } from '../tasks/tasks.controller.js';
function createBrain() {
return {
conversations: {
findAll: vi.fn(),
findById: vi.fn(),
create: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
findMessages: vi.fn(),
addMessage: vi.fn(),
},
projects: {
findAll: vi.fn(),
findAllForUser: vi.fn(),
findById: vi.fn(),
create: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
},
missions: {
findAll: vi.fn(),
findAllByUser: vi.fn(),
findById: vi.fn(),
findByIdAndUser: vi.fn(),
findByProject: vi.fn(),
create: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
},
missionTasks: {
findByMissionAndUser: vi.fn(),
findByIdAndUser: vi.fn(),
create: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
},
tasks: {
findAll: vi.fn(),
findById: vi.fn(),
findByProject: vi.fn(),
findByMission: vi.fn(),
findByStatus: vi.fn(),
create: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
},
};
}
describe('Resource ownership checks', () => {
it('forbids access to another user conversation', async () => {
const brain = createBrain();
brain.conversations.findById.mockResolvedValue({ id: 'conv-1', userId: 'user-2' });
const controller = new ConversationsController(brain as never);
await expect(controller.findOne('conv-1', { id: 'user-1' })).rejects.toBeInstanceOf(
ForbiddenException,
);
});
it('forbids access to another user project', async () => {
const brain = createBrain();
brain.projects.findById.mockResolvedValue({ id: 'project-1', ownerId: 'user-2' });
const teamsService = { canAccessProject: vi.fn().mockResolvedValue(false) };
const controller = new ProjectsController(brain as never, teamsService as never);
await expect(controller.findOne('project-1', { id: 'user-1' })).rejects.toBeInstanceOf(
ForbiddenException,
);
});
it('forbids access to a mission owned by another user', async () => {
const brain = createBrain();
// findByIdAndUser returns undefined when the mission doesn't belong to the user
brain.missions.findByIdAndUser.mockResolvedValue(undefined);
const controller = new MissionsController(brain as never);
await expect(controller.findOne('mission-1', { id: 'user-1' })).rejects.toBeInstanceOf(
NotFoundException,
);
});
it('forbids access to a task owned by another project owner', async () => {
const brain = createBrain();
brain.tasks.findById.mockResolvedValue({ id: 'task-1', projectId: 'project-1' });
brain.projects.findById.mockResolvedValue({ id: 'project-1', ownerId: 'user-2' });
const controller = new TasksController(brain as never);
await expect(controller.findOne('task-1', { id: 'user-1' })).rejects.toBeInstanceOf(
ForbiddenException,
);
});
it('forbids creating a task with an unowned project', async () => {
const brain = createBrain();
brain.projects.findById.mockResolvedValue({ id: 'project-1', ownerId: 'user-2' });
const controller = new TasksController(brain as never);
await expect(
controller.create(
{
title: 'Task',
projectId: 'project-1',
},
{ id: 'user-1' },
),
).rejects.toBeInstanceOf(ForbiddenException);
});
it('forbids listing tasks for an unowned project', async () => {
const brain = createBrain();
brain.projects.findById.mockResolvedValue({ id: 'project-1', ownerId: 'user-2' });
const controller = new TasksController(brain as never);
await expect(
controller.list({ id: 'user-1' }, 'project-1', undefined, undefined),
).rejects.toBeInstanceOf(ForbiddenException);
});
it('lists only tasks for the current user owned projects when no filter is provided', async () => {
const brain = createBrain();
brain.projects.findAll.mockResolvedValue([
{ id: 'project-1', ownerId: 'user-1' },
{ id: 'project-2', ownerId: 'user-2' },
]);
brain.missions.findAll.mockResolvedValue([{ id: 'mission-1', projectId: 'project-1' }]);
brain.tasks.findAll.mockResolvedValue([
{ id: 'task-1', projectId: 'project-1' },
{ id: 'task-2', missionId: 'mission-1' },
{ id: 'task-3', projectId: 'project-2' },
]);
const controller = new TasksController(brain as never);
await expect(
controller.list({ id: 'user-1' }, undefined, undefined, undefined),
).resolves.toEqual([
{ id: 'task-1', projectId: 'project-1' },
{ id: 'task-2', missionId: 'mission-1' },
]);
});
});
@@ -1,73 +0,0 @@
import { Controller, Get, Inject, UseGuards } from '@nestjs/common';
import { sql, type Db } from '@mosaic/db';
import { createQueue } from '@mosaic/queue';
import { DB } from '../database/database.module.js';
import { AgentService } from '../agent/agent.service.js';
import { ProviderService } from '../agent/provider.service.js';
import { AdminGuard } from './admin.guard.js';
import type { HealthStatusDto, ServiceStatusDto } from './admin.dto.js';
@Controller('api/admin/health')
@UseGuards(AdminGuard)
export class AdminHealthController {
constructor(
@Inject(DB) private readonly db: Db,
@Inject(AgentService) private readonly agentService: AgentService,
@Inject(ProviderService) private readonly providerService: ProviderService,
) {}
@Get()
async check(): Promise<HealthStatusDto> {
const [database, cache] = await Promise.all([this.checkDatabase(), this.checkCache()]);
const sessions = this.agentService.listSessions();
const providers = this.providerService.listProviders();
const allOk = database.status === 'ok' && cache.status === 'ok';
return {
status: allOk ? 'ok' : 'degraded',
database,
cache,
agentPool: { activeSessions: sessions.length },
providers: providers.map((p) => ({
id: p.id,
name: p.name,
available: p.available,
modelCount: p.models.length,
})),
checkedAt: new Date().toISOString(),
};
}
private async checkDatabase(): Promise<ServiceStatusDto> {
const start = Date.now();
try {
await this.db.execute(sql`SELECT 1`);
return { status: 'ok', latencyMs: Date.now() - start };
} catch (err) {
return {
status: 'error',
latencyMs: Date.now() - start,
error: err instanceof Error ? err.message : String(err),
};
}
}
private async checkCache(): Promise<ServiceStatusDto> {
const start = Date.now();
const handle = createQueue();
try {
await handle.redis.ping();
return { status: 'ok', latencyMs: Date.now() - start };
} catch (err) {
return {
status: 'error',
latencyMs: Date.now() - start,
error: err instanceof Error ? err.message : String(err),
};
} finally {
await handle.close().catch(() => {});
}
}
}
-146
View File
@@ -1,146 +0,0 @@
import {
Body,
Controller,
Delete,
Get,
HttpCode,
HttpStatus,
Inject,
InternalServerErrorException,
NotFoundException,
Param,
Patch,
Post,
UseGuards,
} from '@nestjs/common';
import { eq, type Db, users as usersTable } from '@mosaic/db';
import type { Auth } from '@mosaic/auth';
import { AUTH } from '../auth/auth.tokens.js';
import { DB } from '../database/database.module.js';
import { AdminGuard } from './admin.guard.js';
import type {
BanUserDto,
CreateUserDto,
UpdateUserRoleDto,
UserDto,
UserListDto,
} from './admin.dto.js';
type UserRow = typeof usersTable.$inferSelect;
function toUserDto(u: UserRow): UserDto {
return {
id: u.id,
name: u.name,
email: u.email,
role: u.role,
banned: u.banned ?? false,
banReason: u.banReason ?? null,
createdAt: u.createdAt.toISOString(),
updatedAt: u.updatedAt.toISOString(),
};
}
async function requireUpdated(
db: Db,
id: string,
update: Partial<Omit<UserRow, 'id' | 'createdAt'>>,
): Promise<UserDto> {
const [updated] = await db
.update(usersTable)
.set({ ...update, updatedAt: new Date() })
.where(eq(usersTable.id, id))
.returning();
if (!updated) throw new InternalServerErrorException('Update returned no rows');
return toUserDto(updated);
}
@Controller('api/admin/users')
@UseGuards(AdminGuard)
export class AdminController {
constructor(
@Inject(DB) private readonly db: Db,
@Inject(AUTH) private readonly auth: Auth,
) {}
@Get()
async listUsers(): Promise<UserListDto> {
const rows = await this.db.select().from(usersTable).orderBy(usersTable.createdAt);
const userList: UserDto[] = rows.map(toUserDto);
return { users: userList, total: userList.length };
}
@Get(':id')
async getUser(@Param('id') id: string): Promise<UserDto> {
const [user] = await this.db.select().from(usersTable).where(eq(usersTable.id, id)).limit(1);
if (!user) throw new NotFoundException('User not found');
return toUserDto(user);
}
@Post()
async createUser(@Body() body: CreateUserDto): Promise<UserDto> {
// Use auth API to create user so password is properly hashed
const authApi = this.auth.api as unknown as {
createUser: (opts: {
body: { name: string; email: string; password: string; role?: string };
}) => Promise<{
user: { id: string; name: string; email: string; createdAt: unknown; updatedAt: unknown };
}>;
};
const result = await authApi.createUser({
body: {
name: body.name,
email: body.email,
password: body.password,
role: body.role ?? 'member',
},
});
// Re-fetch from DB to get full row with our schema
const [user] = await this.db
.select()
.from(usersTable)
.where(eq(usersTable.id, result.user.id))
.limit(1);
if (!user) throw new InternalServerErrorException('User created but not found in DB');
return toUserDto(user);
}
@Patch(':id/role')
async setRole(@Param('id') id: string, @Body() body: UpdateUserRoleDto): Promise<UserDto> {
await this.ensureExists(id);
return requireUpdated(this.db, id, { role: body.role });
}
@Post(':id/ban')
@HttpCode(HttpStatus.OK)
async banUser(@Param('id') id: string, @Body() body: BanUserDto): Promise<UserDto> {
await this.ensureExists(id);
return requireUpdated(this.db, id, { banned: true, banReason: body.reason ?? null });
}
@Post(':id/unban')
@HttpCode(HttpStatus.OK)
async unbanUser(@Param('id') id: string): Promise<UserDto> {
await this.ensureExists(id);
return requireUpdated(this.db, id, { banned: false, banReason: null, banExpires: null });
}
@Delete(':id')
@HttpCode(HttpStatus.NO_CONTENT)
async deleteUser(@Param('id') id: string): Promise<void> {
await this.ensureExists(id);
await this.db.delete(usersTable).where(eq(usersTable.id, id));
}
private async ensureExists(id: string): Promise<void> {
const [existing] = await this.db
.select({ id: usersTable.id })
.from(usersTable)
.where(eq(usersTable.id, id))
.limit(1);
if (!existing) throw new NotFoundException('User not found');
}
}
-56
View File
@@ -1,56 +0,0 @@
export interface UserDto {
id: string;
name: string;
email: string;
role: string;
banned: boolean;
banReason: string | null;
createdAt: string;
updatedAt: string;
}
export interface UserListDto {
users: UserDto[];
total: number;
}
export interface CreateUserDto {
name: string;
email: string;
password: string;
role?: string;
}
export interface UpdateUserRoleDto {
role: string;
}
export interface BanUserDto {
reason?: string;
}
export interface HealthStatusDto {
status: 'ok' | 'degraded' | 'error';
database: ServiceStatusDto;
cache: ServiceStatusDto;
agentPool: AgentPoolStatusDto;
providers: ProviderStatusDto[];
checkedAt: string;
}
export interface ServiceStatusDto {
status: 'ok' | 'error';
latencyMs?: number;
error?: string;
}
export interface AgentPoolStatusDto {
activeSessions: number;
}
export interface ProviderStatusDto {
id: string;
name: string;
available: boolean;
modelCount: number;
}
-64
View File
@@ -1,64 +0,0 @@
import {
CanActivate,
ExecutionContext,
ForbiddenException,
Inject,
Injectable,
UnauthorizedException,
} from '@nestjs/common';
import { fromNodeHeaders } from 'better-auth/node';
import type { Auth } from '@mosaic/auth';
import type { Db } from '@mosaic/db';
import { eq, users as usersTable } from '@mosaic/db';
import type { FastifyRequest } from 'fastify';
import { AUTH } from '../auth/auth.tokens.js';
import { DB } from '../database/database.module.js';
interface UserWithRole {
id: string;
role?: string;
}
@Injectable()
export class AdminGuard implements CanActivate {
constructor(
@Inject(AUTH) private readonly auth: Auth,
@Inject(DB) private readonly db: Db,
) {}
async canActivate(context: ExecutionContext): Promise<boolean> {
const request = context.switchToHttp().getRequest<FastifyRequest>();
const headers = fromNodeHeaders(request.raw.headers);
const result = await this.auth.api.getSession({ headers });
if (!result) {
throw new UnauthorizedException('Invalid or expired session');
}
const user = result.user as UserWithRole;
// Ensure the role field is populated. better-auth should include additionalFields
// in the session, but as a fallback, fetch the role from the database if needed.
let userRole = user.role;
if (!userRole) {
const [dbUser] = await this.db
.select({ role: usersTable.role })
.from(usersTable)
.where(eq(usersTable.id, user.id))
.limit(1);
userRole = dbUser?.role ?? 'member';
// Update the session user object with the fetched role
(user as UserWithRole).role = userRole;
}
if (userRole !== 'admin') {
throw new ForbiddenException('Admin access required');
}
(request as FastifyRequest & { user: unknown; session: unknown }).user = result.user;
(request as FastifyRequest & { user: unknown; session: unknown }).session = result.session;
return true;
}
}
-10
View File
@@ -1,10 +0,0 @@
import { Module } from '@nestjs/common';
import { AdminController } from './admin.controller.js';
import { AdminHealthController } from './admin-health.controller.js';
import { AdminGuard } from './admin.guard.js';
@Module({
controllers: [AdminController, AdminHealthController],
providers: [AdminGuard],
})
export class AdminModule {}
@@ -1,138 +0,0 @@
import { describe, it, expect, beforeEach, vi } from 'vitest';
import { RoutingService } from '../routing.service.js';
import type { ModelInfo } from '@mosaic/types';
const mockModels: ModelInfo[] = [
{
id: 'claude-3-haiku',
provider: 'anthropic',
name: 'Claude 3 Haiku',
reasoning: false,
contextWindow: 200_000,
maxTokens: 4096,
inputTypes: ['text', 'image'],
cost: { input: 0.25, output: 1.25, cacheRead: 0.03, cacheWrite: 0.3 },
},
{
id: 'claude-3-sonnet',
provider: 'anthropic',
name: 'Claude 3 Sonnet',
reasoning: true,
contextWindow: 200_000,
maxTokens: 8192,
inputTypes: ['text', 'image'],
cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 },
},
{
id: 'llama3.2',
provider: 'ollama',
name: 'Llama 3.2',
reasoning: false,
contextWindow: 128_000,
maxTokens: 4096,
inputTypes: ['text'],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
},
];
function createMockProviderService() {
return {
listAvailableModels: vi.fn().mockReturnValue(mockModels),
findModel: vi.fn(),
getDefaultModel: vi.fn(),
getRegistry: vi.fn(),
listProviders: vi.fn(),
registerCustomProvider: vi.fn(),
};
}
describe('RoutingService', () => {
let routingService: RoutingService;
let mockProviderService: ReturnType<typeof createMockProviderService>;
beforeEach(() => {
mockProviderService = createMockProviderService();
routingService = new RoutingService(mockProviderService as never);
});
it('returns a model when no criteria specified', () => {
const result = routingService.route();
expect(result).not.toBeNull();
expect(result!.provider).toBeDefined();
expect(result!.modelId).toBeDefined();
expect(result!.score).toBeGreaterThan(0);
});
it('returns null when no models available', () => {
mockProviderService.listAvailableModels.mockReturnValue([]);
const result = routingService.route();
expect(result).toBeNull();
});
it('selects preferred model when specified', () => {
const result = routingService.route({
preferredProvider: 'anthropic',
preferredModel: 'claude-3-sonnet',
});
expect(result).not.toBeNull();
expect(result!.provider).toBe('anthropic');
expect(result!.modelId).toBe('claude-3-sonnet');
expect(result!.score).toBe(100);
});
it('disqualifies models without reasoning when required', () => {
const result = routingService.route({ requireReasoning: true });
expect(result).not.toBeNull();
expect(result!.modelId).toBe('claude-3-sonnet');
});
it('disqualifies models without image input when required', () => {
const result = routingService.route({ requireImageInput: true });
expect(result).not.toBeNull();
// Llama doesn't support images, should be excluded
expect(result!.provider).toBe('anthropic');
});
it('respects minimum context window', () => {
const result = routingService.route({ minContextWindow: 150_000 });
expect(result).not.toBeNull();
// Only anthropic models have 200k context
expect(result!.provider).toBe('anthropic');
});
it('favors cheap models when costTier is cheap', () => {
const result = routingService.route({ costTier: 'cheap' });
expect(result).not.toBeNull();
// Ollama (free) and Haiku ($0.25/M) are cheap
expect(['ollama', 'anthropic'].includes(result!.provider)).toBe(true);
if (result!.provider === 'anthropic') {
expect(result!.modelId).toBe('claude-3-haiku');
}
});
it('ranks all models and returns sorted results', () => {
const ranked = routingService.rank({ taskType: 'coding' });
expect(ranked.length).toBeGreaterThan(0);
// Should be sorted by score descending
for (let i = 1; i < ranked.length; i++) {
expect(ranked[i]!.score).toBeLessThanOrEqual(ranked[i - 1]!.score);
}
});
it('gives reasoning bonus for coding tasks', () => {
const ranked = routingService.rank({ taskType: 'coding' });
const sonnet = ranked.find((r) => r.modelId === 'claude-3-sonnet');
const haiku = ranked.find((r) => r.modelId === 'claude-3-haiku');
expect(sonnet).toBeDefined();
expect(haiku).toBeDefined();
// Sonnet (reasoning) should score higher for coding than haiku (no reasoning)
expect(sonnet!.score).toBeGreaterThan(haiku!.score);
});
it('prefers specified provider', () => {
const ranked = routingService.rank({ preferredProvider: 'ollama' });
const ollamaModel = ranked.find((r) => r.provider === 'ollama');
expect(ollamaModel).toBeDefined();
expect(ollamaModel!.reasoning).toContain('preferred provider');
});
});
@@ -1,97 +0,0 @@
import {
IsArray,
IsBoolean,
IsIn,
IsObject,
IsOptional,
IsString,
IsUUID,
MaxLength,
} from 'class-validator';
const agentStatuses = ['idle', 'active', 'error', 'offline'] as const;
export class CreateAgentConfigDto {
@IsString()
@MaxLength(255)
name!: string;
@IsString()
@MaxLength(255)
provider!: string;
@IsString()
@MaxLength(255)
model!: string;
@IsOptional()
@IsIn(agentStatuses)
status?: 'idle' | 'active' | 'error' | 'offline';
@IsOptional()
@IsUUID()
projectId?: string;
@IsOptional()
@IsString()
@MaxLength(50_000)
systemPrompt?: string;
@IsOptional()
@IsArray()
allowedTools?: string[];
@IsOptional()
@IsArray()
skills?: string[];
@IsOptional()
@IsBoolean()
isSystem?: boolean;
@IsOptional()
@IsObject()
config?: Record<string, unknown>;
}
export class UpdateAgentConfigDto {
@IsOptional()
@IsString()
@MaxLength(255)
name?: string;
@IsOptional()
@IsString()
@MaxLength(255)
provider?: string;
@IsOptional()
@IsString()
@MaxLength(255)
model?: string;
@IsOptional()
@IsIn(agentStatuses)
status?: 'idle' | 'active' | 'error' | 'offline';
@IsOptional()
@IsUUID()
projectId?: string | null;
@IsOptional()
@IsString()
@MaxLength(50_000)
systemPrompt?: string | null;
@IsOptional()
@IsArray()
allowedTools?: string[] | null;
@IsOptional()
@IsArray()
skills?: string[] | null;
@IsOptional()
@IsObject()
config?: Record<string, unknown> | null;
}
@@ -1,84 +0,0 @@
import {
Body,
Controller,
Delete,
ForbiddenException,
Get,
HttpCode,
HttpStatus,
Inject,
NotFoundException,
Param,
Patch,
Post,
UseGuards,
} from '@nestjs/common';
import type { Brain } from '@mosaic/brain';
import { BRAIN } from '../brain/brain.tokens.js';
import { AuthGuard } from '../auth/auth.guard.js';
import { CurrentUser } from '../auth/current-user.decorator.js';
import { CreateAgentConfigDto, UpdateAgentConfigDto } from './agent-config.dto.js';
@Controller('api/agents')
@UseGuards(AuthGuard)
export class AgentConfigsController {
constructor(@Inject(BRAIN) private readonly brain: Brain) {}
@Get()
async list(@CurrentUser() user: { id: string; role?: string }) {
return this.brain.agents.findAccessible(user.id);
}
@Get(':id')
async findOne(@Param('id') id: string, @CurrentUser() user: { id: string }) {
const agent = await this.brain.agents.findById(id);
if (!agent) throw new NotFoundException('Agent not found');
if (!agent.isSystem && agent.ownerId !== user.id) {
throw new ForbiddenException('Agent does not belong to the current user');
}
return agent;
}
@Post()
async create(@Body() dto: CreateAgentConfigDto, @CurrentUser() user: { id: string }) {
return this.brain.agents.create({
...dto,
ownerId: user.id,
isSystem: false,
});
}
@Patch(':id')
async update(
@Param('id') id: string,
@Body() dto: UpdateAgentConfigDto,
@CurrentUser() user: { id: string; role?: string },
) {
const agent = await this.brain.agents.findById(id);
if (!agent) throw new NotFoundException('Agent not found');
if (agent.isSystem && user.role !== 'admin') {
throw new ForbiddenException('Only admins can update system agents');
}
if (!agent.isSystem && agent.ownerId !== user.id) {
throw new ForbiddenException('Agent does not belong to the current user');
}
const updated = await this.brain.agents.update(id, dto);
if (!updated) throw new NotFoundException('Agent not found');
return updated;
}
@Delete(':id')
@HttpCode(HttpStatus.NO_CONTENT)
async remove(@Param('id') id: string, @CurrentUser() user: { id: string; role?: string }) {
const agent = await this.brain.agents.findById(id);
if (!agent) throw new NotFoundException('Agent not found');
if (agent.isSystem) {
throw new ForbiddenException('Cannot delete system agents');
}
if (agent.ownerId !== user.id) {
throw new ForbiddenException('Agent does not belong to the current user');
}
const deleted = await this.brain.agents.remove(id);
if (!deleted) throw new NotFoundException('Agent not found');
}
}
-21
View File
@@ -1,21 +0,0 @@
import { Global, Module } from '@nestjs/common';
import { AgentService } from './agent.service.js';
import { ProviderService } from './provider.service.js';
import { RoutingService } from './routing.service.js';
import { SkillLoaderService } from './skill-loader.service.js';
import { ProvidersController } from './providers.controller.js';
import { SessionsController } from './sessions.controller.js';
import { AgentConfigsController } from './agent-configs.controller.js';
import { CoordModule } from '../coord/coord.module.js';
import { McpClientModule } from '../mcp-client/mcp-client.module.js';
import { SkillsModule } from '../skills/skills.module.js';
import { GCModule } from '../gc/gc.module.js';
@Global()
@Module({
imports: [CoordModule, McpClientModule, SkillsModule, GCModule],
providers: [ProviderService, RoutingService, SkillLoaderService, AgentService],
controllers: [ProvidersController, SessionsController, AgentConfigsController],
exports: [AgentService, ProviderService, RoutingService, SkillLoaderService],
})
export class AgentModule {}
-455
View File
@@ -1,455 +0,0 @@
import { Inject, Injectable, Logger, Optional, type OnModuleDestroy } from '@nestjs/common';
import {
createAgentSession,
DefaultResourceLoader,
SessionManager,
type AgentSession as PiAgentSession,
type AgentSessionEvent,
type ToolDefinition,
} from '@mariozechner/pi-coding-agent';
import type { Brain } from '@mosaic/brain';
import type { Memory } from '@mosaic/memory';
import { BRAIN } from '../brain/brain.tokens.js';
import { MEMORY } from '../memory/memory.tokens.js';
import { EmbeddingService } from '../memory/embedding.service.js';
import { CoordService } from '../coord/coord.service.js';
import { ProviderService } from './provider.service.js';
import { McpClientService } from '../mcp-client/mcp-client.service.js';
import { SkillLoaderService } from './skill-loader.service.js';
import { createBrainTools } from './tools/brain-tools.js';
import { createCoordTools } from './tools/coord-tools.js';
import { createMemoryTools } from './tools/memory-tools.js';
import { createFileTools } from './tools/file-tools.js';
import { createGitTools } from './tools/git-tools.js';
import { createShellTools } from './tools/shell-tools.js';
import { createWebTools } from './tools/web-tools.js';
import type { SessionInfoDto } from './session.dto.js';
import { SystemOverrideService } from '../preferences/system-override.service.js';
import { PreferencesService } from '../preferences/preferences.service.js';
import { SessionGCService } from '../gc/session-gc.service.js';
export interface AgentSessionOptions {
provider?: string;
modelId?: string;
/**
* Sandbox working directory for the session.
* File, git, and shell tools will be restricted to this directory.
* Falls back to AGENT_FILE_SANDBOX_DIR env var or process.cwd().
*/
sandboxDir?: string;
/**
* Platform-level system prompt for this session.
* Merged with skill prompt additions (platform prompt first, then skills).
* Falls back to AGENT_SYSTEM_PROMPT env var when omitted.
*/
systemPrompt?: string;
/**
* Explicit allowlist of tool names available in this session.
* When set, only listed tools are registered with the agent.
* When omitted for non-admin users, falls back to AGENT_USER_TOOLS env var.
* Admins (isAdmin=true) always receive the full tool set unless explicitly restricted.
*/
allowedTools?: string[];
/** Whether the requesting user has admin privileges. Controls default tool access. */
isAdmin?: boolean;
/**
* DB agent config ID. When provided, loads agent config from DB and merges
* provider, model, systemPrompt, and allowedTools. Explicit call-site options
* take precedence over config values.
*/
agentConfigId?: string;
/** ID of the user who owns this session. Used for preferences and system override lookups. */
userId?: string;
}
export interface AgentSession {
id: string;
provider: string;
modelId: string;
piSession: PiAgentSession;
listeners: Set<(event: AgentSessionEvent) => void>;
unsubscribe: () => void;
createdAt: number;
promptCount: number;
channels: Set<string>;
/** System prompt additions injected from enabled prompt-type skills. */
skillPromptAdditions: string[];
/** Resolved sandbox directory for this session. */
sandboxDir: string;
/** Tool names available in this session, or null when all tools are available. */
allowedTools: string[] | null;
/** User ID that owns this session, used for preference lookups. */
userId?: string;
}
@Injectable()
export class AgentService implements OnModuleDestroy {
private readonly logger = new Logger(AgentService.name);
private readonly sessions = new Map<string, AgentSession>();
private readonly creating = new Map<string, Promise<AgentSession>>();
constructor(
@Inject(ProviderService) private readonly providerService: ProviderService,
@Inject(BRAIN) private readonly brain: Brain,
@Inject(MEMORY) private readonly memory: Memory,
@Inject(EmbeddingService) private readonly embeddingService: EmbeddingService,
@Inject(CoordService) private readonly coordService: CoordService,
@Inject(McpClientService) private readonly mcpClientService: McpClientService,
@Inject(SkillLoaderService) private readonly skillLoaderService: SkillLoaderService,
@Optional()
@Inject(SystemOverrideService)
private readonly systemOverride: SystemOverrideService | null,
@Optional()
@Inject(PreferencesService)
private readonly preferencesService: PreferencesService | null,
@Inject(SessionGCService) private readonly gc: SessionGCService,
) {}
/**
* Build the full set of custom tools scoped to the given sandbox directory.
* Brain/coord/memory/web tools are stateless with respect to cwd; file/git/shell
* tools receive the resolved sandboxDir so they operate within the sandbox.
*/
private buildToolsForSandbox(sandboxDir: string): ToolDefinition[] {
return [
...createBrainTools(this.brain),
...createCoordTools(this.coordService),
...createMemoryTools(
this.memory,
this.embeddingService.available ? this.embeddingService : null,
),
...createFileTools(sandboxDir),
...createGitTools(sandboxDir),
...createShellTools(sandboxDir),
...createWebTools(),
];
}
/**
* Resolve the tool allowlist for a session.
* - Admin users: all tools unless an explicit allowedTools list is passed.
* - Regular users: use allowedTools if provided, otherwise parse AGENT_USER_TOOLS env var.
* Returns null when all tools should be available.
*/
private resolveAllowedTools(isAdmin: boolean, allowedTools?: string[]): string[] | null {
if (allowedTools !== undefined) {
return allowedTools.length === 0 ? [] : allowedTools;
}
if (isAdmin) {
return null; // admins get everything
}
const envTools = process.env['AGENT_USER_TOOLS'];
if (!envTools) {
return null; // no restriction configured
}
return envTools
.split(',')
.map((t) => t.trim())
.filter((t) => t.length > 0);
}
async createSession(sessionId: string, options?: AgentSessionOptions): Promise<AgentSession> {
const existing = this.sessions.get(sessionId);
if (existing) return existing;
const inflight = this.creating.get(sessionId);
if (inflight) return inflight;
const promise = this.doCreateSession(sessionId, options).finally(() => {
this.creating.delete(sessionId);
});
this.creating.set(sessionId, promise);
return promise;
}
private async doCreateSession(
sessionId: string,
options?: AgentSessionOptions,
): Promise<AgentSession> {
// Merge DB agent config when agentConfigId is provided
let mergedOptions = options;
if (options?.agentConfigId) {
const agentConfig = await this.brain.agents.findById(options.agentConfigId);
if (agentConfig) {
mergedOptions = {
provider: options.provider ?? agentConfig.provider,
modelId: options.modelId ?? agentConfig.model,
systemPrompt: options.systemPrompt ?? agentConfig.systemPrompt ?? undefined,
allowedTools: options.allowedTools ?? agentConfig.allowedTools ?? undefined,
sandboxDir: options.sandboxDir,
isAdmin: options.isAdmin,
agentConfigId: options.agentConfigId,
};
this.logger.log(
`Merged agent config "${agentConfig.name}" (${agentConfig.id}) into session ${sessionId}`,
);
}
}
const model = this.resolveModel(mergedOptions);
const providerName = model?.provider ?? 'default';
const modelId = model?.id ?? 'default';
// Resolve sandbox directory: option > env var > process.cwd()
const sandboxDir =
mergedOptions?.sandboxDir ?? process.env['AGENT_FILE_SANDBOX_DIR'] ?? process.cwd();
// Resolve allowed tool set
const allowedTools = this.resolveAllowedTools(
mergedOptions?.isAdmin ?? false,
mergedOptions?.allowedTools,
);
this.logger.log(
`Creating agent session: ${sessionId} (provider=${providerName}, model=${modelId}, sandbox=${sandboxDir}, tools=${allowedTools === null ? 'all' : allowedTools.join(',') || 'none'})`,
);
// Load skill tools from the catalog
const { metaTools: skillMetaTools, promptAdditions } =
await this.skillLoaderService.loadForSession();
if (skillMetaTools.length > 0) {
this.logger.log(`Attaching ${skillMetaTools.length} skill tool(s) to session ${sessionId}`);
}
if (promptAdditions.length > 0) {
this.logger.log(
`Injecting ${promptAdditions.length} skill prompt addition(s) into session ${sessionId}`,
);
}
// Build per-session tools scoped to the sandbox directory
const sandboxTools = this.buildToolsForSandbox(sandboxDir);
// Combine static tools with dynamically discovered MCP client tools and skill tools
const mcpTools = this.mcpClientService.getToolDefinitions();
let allCustomTools = [...sandboxTools, ...skillMetaTools, ...mcpTools];
if (mcpTools.length > 0) {
this.logger.log(`Attaching ${mcpTools.length} MCP client tool(s) to session ${sessionId}`);
}
// Filter tools by allowlist when a restriction is in effect
if (allowedTools !== null) {
const allowedSet = new Set(allowedTools);
const before = allCustomTools.length;
allCustomTools = allCustomTools.filter((t) => allowedSet.has(t.name));
this.logger.log(
`Tool restriction applied: ${allCustomTools.length}/${before} tools allowed for session ${sessionId}`,
);
}
// Build system prompt: platform prompt + skill additions appended
const platformPrompt =
mergedOptions?.systemPrompt ?? process.env['AGENT_SYSTEM_PROMPT'] ?? undefined;
const appendSystemPrompt =
promptAdditions.length > 0 ? promptAdditions.join('\n\n') : undefined;
// Construct a resource loader that injects the configured system prompt
const resourceLoader = new DefaultResourceLoader({
cwd: sandboxDir,
noExtensions: true,
noSkills: true,
noPromptTemplates: true,
noThemes: true,
systemPrompt: platformPrompt,
appendSystemPrompt: appendSystemPrompt,
});
await resourceLoader.reload();
let piSession: PiAgentSession;
try {
const result = await createAgentSession({
sessionManager: SessionManager.inMemory(),
modelRegistry: this.providerService.getRegistry(),
model: model ?? undefined,
cwd: sandboxDir,
tools: [],
customTools: allCustomTools,
resourceLoader,
});
piSession = result.session;
} catch (err) {
this.logger.error(
`Failed to create agent session for ${sessionId}`,
err instanceof Error ? err.stack : String(err),
);
throw new Error(`Agent session creation failed for ${sessionId}: ${String(err)}`);
}
const listeners = new Set<(event: AgentSessionEvent) => void>();
const unsubscribe = piSession.subscribe((event) => {
for (const listener of listeners) {
try {
listener(event);
} catch (err) {
this.logger.error(`Event listener error in session ${sessionId}`, err);
}
}
});
const session: AgentSession = {
id: sessionId,
provider: providerName,
modelId,
piSession,
listeners,
unsubscribe,
createdAt: Date.now(),
promptCount: 0,
channels: new Set(),
skillPromptAdditions: promptAdditions,
sandboxDir,
allowedTools,
userId: mergedOptions?.userId,
};
this.sessions.set(sessionId, session);
this.logger.log(`Agent session ${sessionId} ready (${providerName}/${modelId})`);
return session;
}
private resolveModel(options?: AgentSessionOptions) {
if (!options?.provider && !options?.modelId) {
return this.providerService.getDefaultModel() ?? null;
}
if (options.provider && options.modelId) {
const model = this.providerService.findModel(options.provider, options.modelId);
if (!model) {
throw new Error(`Model not found: ${options.provider}/${options.modelId}`);
}
return model;
}
if (options.modelId) {
const available = this.providerService.listAvailableModels();
const match = available.find((m) => m.id === options.modelId);
if (match) {
return this.providerService.findModel(match.provider, match.id) ?? null;
}
}
return this.providerService.getDefaultModel() ?? null;
}
getSession(sessionId: string): AgentSession | undefined {
return this.sessions.get(sessionId);
}
listSessions(): SessionInfoDto[] {
const now = Date.now();
return Array.from(this.sessions.values()).map((s) => ({
id: s.id,
provider: s.provider,
modelId: s.modelId,
createdAt: new Date(s.createdAt).toISOString(),
promptCount: s.promptCount,
channels: Array.from(s.channels),
durationMs: now - s.createdAt,
}));
}
getSessionInfo(sessionId: string): SessionInfoDto | undefined {
const s = this.sessions.get(sessionId);
if (!s) return undefined;
return {
id: s.id,
provider: s.provider,
modelId: s.modelId,
createdAt: new Date(s.createdAt).toISOString(),
promptCount: s.promptCount,
channels: Array.from(s.channels),
durationMs: Date.now() - s.createdAt,
};
}
addChannel(sessionId: string, channel: string): void {
const session = this.sessions.get(sessionId);
if (session) {
session.channels.add(channel);
}
}
removeChannel(sessionId: string, channel: string): void {
const session = this.sessions.get(sessionId);
if (session) {
session.channels.delete(channel);
}
}
async prompt(sessionId: string, message: string): Promise<void> {
const session = this.sessions.get(sessionId);
if (!session) {
throw new Error(`No agent session found: ${sessionId}`);
}
session.promptCount += 1;
// Prepend session-scoped system override if present (renew TTL on each turn)
let effectiveMessage = message;
if (this.systemOverride) {
const override = await this.systemOverride.get(sessionId);
if (override) {
effectiveMessage = `[System Override]\n${override}\n\n${message}`;
await this.systemOverride.renew(sessionId);
this.logger.debug(`Applied system override for session ${sessionId}`);
}
}
try {
await session.piSession.prompt(effectiveMessage);
} catch (err) {
this.logger.error(
`Prompt failed for session=${sessionId}, messageLength=${message.length}`,
err instanceof Error ? err.stack : String(err),
);
throw err;
}
}
onEvent(sessionId: string, listener: (event: AgentSessionEvent) => void): () => void {
const session = this.sessions.get(sessionId);
if (!session) {
throw new Error(`No agent session found: ${sessionId}`);
}
session.listeners.add(listener);
return () => session.listeners.delete(listener);
}
async destroySession(sessionId: string): Promise<void> {
const session = this.sessions.get(sessionId);
if (!session) return;
this.logger.log(`Destroying agent session ${sessionId}`);
try {
session.unsubscribe();
} catch (err) {
this.logger.error(`Failed to unsubscribe session ${sessionId}`, String(err));
}
try {
session.piSession.dispose();
} catch (err) {
this.logger.error(`Failed to dispose piSession for ${sessionId}`, String(err));
}
session.listeners.clear();
session.channels.clear();
this.sessions.delete(sessionId);
// Run GC cleanup for this session (fire and forget, errors are logged)
this.gc.collect(sessionId).catch((err: unknown) => {
this.logger.error(
`GC collect failed for session ${sessionId}`,
err instanceof Error ? err.stack : String(err),
);
});
}
async onModuleDestroy(): Promise<void> {
this.logger.log('Shutting down all agent sessions');
const stops = Array.from(this.sessions.keys()).map((id) => this.destroySession(id));
const results = await Promise.allSettled(stops);
for (const result of results) {
if (result.status === 'rejected') {
this.logger.error('Session shutdown failure', String(result.reason));
}
}
}
}
-17
View File
@@ -1,17 +0,0 @@
export interface TestConnectionDto {
/** Provider identifier to test (e.g. 'ollama', custom provider id) */
providerId: string;
/** Optional base URL override for ad-hoc testing */
baseUrl?: string;
}
export interface TestConnectionResultDto {
providerId: string;
reachable: boolean;
/** Round-trip latency in milliseconds (present when reachable) */
latencyMs?: number;
/** Human-readable error when unreachable */
error?: string;
/** Model ids discovered at the remote endpoint (present when reachable) */
discoveredModels?: string[];
}
-199
View File
@@ -1,199 +0,0 @@
import { Injectable, Logger, type OnModuleInit } from '@nestjs/common';
import { ModelRegistry, AuthStorage } from '@mariozechner/pi-coding-agent';
import type { Model, Api } from '@mariozechner/pi-ai';
import type { ModelInfo, ProviderInfo, CustomProviderConfig } from '@mosaic/types';
import type { TestConnectionResultDto } from './provider.dto.js';
@Injectable()
export class ProviderService implements OnModuleInit {
private readonly logger = new Logger(ProviderService.name);
private registry!: ModelRegistry;
onModuleInit(): void {
const authStorage = AuthStorage.inMemory();
this.registry = new ModelRegistry(authStorage);
this.registerOllamaProvider();
this.registerCustomProviders();
const available = this.registry.getAvailable();
this.logger.log(`Providers initialized: ${available.length} models available`);
}
getRegistry(): ModelRegistry {
return this.registry;
}
findModel(provider: string, modelId: string): Model<Api> | undefined {
return this.registry.find(provider, modelId);
}
getDefaultModel(): Model<Api> | undefined {
const available = this.registry.getAvailable();
return available[0];
}
listProviders(): ProviderInfo[] {
const allModels = this.registry.getAll();
const availableModels = this.registry.getAvailable();
const availableIds = new Set(availableModels.map((m) => `${m.provider}:${m.id}`));
const providerMap = new Map<string, ProviderInfo>();
for (const model of allModels) {
let info = providerMap.get(model.provider);
if (!info) {
info = {
id: model.provider,
name: model.provider,
available: false,
models: [],
};
providerMap.set(model.provider, info);
}
const isAvailable = availableIds.has(`${model.provider}:${model.id}`);
if (isAvailable) info.available = true;
info.models.push(this.toModelInfo(model));
}
return Array.from(providerMap.values());
}
listAvailableModels(): ModelInfo[] {
return this.registry.getAvailable().map((m) => this.toModelInfo(m));
}
async testConnection(providerId: string, baseUrl?: string): Promise<TestConnectionResultDto> {
// Resolve baseUrl: explicit override > registered provider > ollama env
let resolvedUrl = baseUrl;
if (!resolvedUrl) {
const allModels = this.registry.getAll();
const providerModels = allModels.filter((m) => m.provider === providerId);
if (providerModels.length === 0) {
return { providerId, reachable: false, error: `Provider '${providerId}' not found` };
}
// For Ollama, derive the base URL from environment
if (providerId === 'ollama') {
const ollamaUrl = process.env['OLLAMA_BASE_URL'] ?? process.env['OLLAMA_HOST'];
if (!ollamaUrl) {
return { providerId, reachable: false, error: 'OLLAMA_BASE_URL not configured' };
}
resolvedUrl = `${ollamaUrl}/v1/models`;
} else {
// For other providers, we can only do a basic check
return { providerId, reachable: true, discoveredModels: providerModels.map((m) => m.id) };
}
} else {
resolvedUrl = resolvedUrl.replace(/\/?$/, '') + '/models';
}
const start = Date.now();
try {
const res = await fetch(resolvedUrl, {
method: 'GET',
headers: { Accept: 'application/json' },
signal: AbortSignal.timeout(5000),
});
const latencyMs = Date.now() - start;
if (!res.ok) {
return { providerId, reachable: false, latencyMs, error: `HTTP ${res.status}` };
}
let discoveredModels: string[] | undefined;
try {
const json = (await res.json()) as { models?: Array<{ id?: string; name?: string }> };
if (Array.isArray(json.models)) {
discoveredModels = json.models.map((m) => m.id ?? m.name ?? '').filter(Boolean);
}
} catch {
// ignore parse errors — endpoint was reachable
}
return { providerId, reachable: true, latencyMs, discoveredModels };
} catch (err) {
const latencyMs = Date.now() - start;
const message = err instanceof Error ? err.message : String(err);
return { providerId, reachable: false, latencyMs, error: message };
}
}
registerCustomProvider(config: CustomProviderConfig): void {
this.registry.registerProvider(config.id, {
baseUrl: config.baseUrl,
apiKey: config.apiKey,
models: config.models.map((m) => ({
id: m.id,
name: m.name,
reasoning: m.reasoning ?? false,
input: ['text'] as ('text' | 'image')[],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
contextWindow: m.contextWindow ?? 4096,
maxTokens: m.maxTokens ?? 4096,
})),
});
this.logger.log(`Registered custom provider: ${config.id} (${config.models.length} models)`);
}
private registerOllamaProvider(): void {
const ollamaUrl = process.env['OLLAMA_BASE_URL'] ?? process.env['OLLAMA_HOST'];
if (!ollamaUrl) return;
const modelsEnv = process.env['OLLAMA_MODELS'] ?? 'llama3.2,codellama,mistral';
const modelIds = modelsEnv
.split(',')
.map((modelId: string) => modelId.trim())
.filter(Boolean);
this.registry.registerProvider('ollama', {
baseUrl: `${ollamaUrl}/v1`,
apiKey: 'ollama',
api: 'openai-completions' as never,
models: modelIds.map((id) => ({
id,
name: id,
reasoning: false,
input: ['text'] as ('text' | 'image')[],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
contextWindow: 8192,
maxTokens: 4096,
})),
});
this.logger.log(
`Ollama provider registered at ${ollamaUrl} with models: ${modelIds.join(', ')}`,
);
}
private registerCustomProviders(): void {
const customJson = process.env['MOSAIC_CUSTOM_PROVIDERS'];
if (!customJson) return;
try {
const configs = JSON.parse(customJson) as CustomProviderConfig[];
for (const config of configs) {
this.registerCustomProvider(config);
}
} catch (err) {
this.logger.error('Failed to parse MOSAIC_CUSTOM_PROVIDERS', String(err));
}
}
private toModelInfo(model: Model<Api>): ModelInfo {
return {
id: model.id,
provider: model.provider,
name: model.name,
reasoning: model.reasoning,
contextWindow: model.contextWindow,
maxTokens: model.maxTokens,
inputTypes: model.input,
cost: model.cost,
};
}
}
@@ -1,40 +0,0 @@
import { Body, Controller, Get, Inject, Post, UseGuards } from '@nestjs/common';
import type { RoutingCriteria } from '@mosaic/types';
import { AuthGuard } from '../auth/auth.guard.js';
import { ProviderService } from './provider.service.js';
import { RoutingService } from './routing.service.js';
import type { TestConnectionDto, TestConnectionResultDto } from './provider.dto.js';
@Controller('api/providers')
@UseGuards(AuthGuard)
export class ProvidersController {
constructor(
@Inject(ProviderService) private readonly providerService: ProviderService,
@Inject(RoutingService) private readonly routingService: RoutingService,
) {}
@Get()
list() {
return this.providerService.listProviders();
}
@Get('models')
listModels() {
return this.providerService.listAvailableModels();
}
@Post('test')
testConnection(@Body() body: TestConnectionDto): Promise<TestConnectionResultDto> {
return this.providerService.testConnection(body.providerId, body.baseUrl);
}
@Post('route')
route(@Body() criteria: RoutingCriteria) {
return this.routingService.route(criteria);
}
@Post('rank')
rank(@Body() criteria: RoutingCriteria) {
return this.routingService.rank(criteria);
}
}
-165
View File
@@ -1,165 +0,0 @@
import { Inject, Injectable, Logger } from '@nestjs/common';
import type { ModelInfo } from '@mosaic/types';
import type { RoutingCriteria, RoutingResult, CostTier } from '@mosaic/types';
import { ProviderService } from './provider.service.js';
/** Per-million-token cost thresholds for tier classification */
const COST_TIER_THRESHOLDS: Record<CostTier, { maxInput: number }> = {
cheap: { maxInput: 1 },
standard: { maxInput: 10 },
premium: { maxInput: Infinity },
};
@Injectable()
export class RoutingService {
private readonly logger = new Logger(RoutingService.name);
constructor(@Inject(ProviderService) private readonly providerService: ProviderService) {}
/**
* Select the best available model for the given criteria.
* Returns null if no model matches the requirements.
*/
route(criteria: RoutingCriteria = {}): RoutingResult | null {
const available = this.providerService.listAvailableModels();
if (available.length === 0) {
this.logger.warn('No available models for routing');
return null;
}
// If a specific model is preferred, try it first
if (criteria.preferredProvider && criteria.preferredModel) {
const match = available.find(
(m) => m.provider === criteria.preferredProvider && m.id === criteria.preferredModel,
);
if (match) {
return {
provider: match.provider,
modelId: match.id,
modelName: match.name,
score: 100,
reasoning: 'Preferred model selected',
};
}
}
// Score and rank candidates
const scored = available
.map((model) => this.scoreModel(model, criteria))
.filter((s) => s.score > 0)
.sort((a, b) => b.score - a.score);
if (scored.length === 0) {
this.logger.warn('No models matched routing criteria', criteria);
return null;
}
const best = scored[0] as RoutingResult;
this.logger.debug(
`Routed to ${best.provider}/${best.modelId} (score=${best.score}): ${best.reasoning}`,
);
return best;
}
/**
* List all available models ranked by suitability for the given criteria.
*/
rank(criteria: RoutingCriteria = {}): RoutingResult[] {
const available = this.providerService.listAvailableModels();
return available
.map((model) => this.scoreModel(model, criteria))
.filter((s) => s.score > 0)
.sort((a, b) => b.score - a.score);
}
private scoreModel(model: ModelInfo, criteria: RoutingCriteria): RoutingResult {
let score = 50; // Base score
const reasons: string[] = [];
// Hard requirements — disqualify if not met
if (criteria.requireReasoning && !model.reasoning) {
return this.disqualified(model, 'reasoning required but not supported');
}
if (criteria.requireImageInput && !model.inputTypes.includes('image')) {
return this.disqualified(model, 'image input required but not supported');
}
if (criteria.minContextWindow && model.contextWindow < criteria.minContextWindow) {
return this.disqualified(
model,
`context window ${model.contextWindow} < required ${criteria.minContextWindow}`,
);
}
// Cost tier matching
if (criteria.costTier) {
const tier = this.classifyTier(model);
if (tier === criteria.costTier) {
score += 20;
reasons.push(`cost tier match (${tier})`);
} else if (
(criteria.costTier === 'cheap' && tier === 'standard') ||
(criteria.costTier === 'standard' && tier === 'premium')
) {
score += 5;
reasons.push(`adjacent cost tier (wanted ${criteria.costTier}, got ${tier})`);
}
}
// Prefer cheaper models when no cost tier specified
if (!criteria.costTier) {
const costPerMillion = model.cost.input;
if (costPerMillion <= 1) score += 10;
else if (costPerMillion <= 5) score += 5;
}
// Provider preference
if (criteria.preferredProvider && model.provider === criteria.preferredProvider) {
score += 15;
reasons.push('preferred provider');
}
// Reasoning bonus for complex tasks
if (model.reasoning) {
if (criteria.taskType === 'coding' || criteria.taskType === 'analysis') {
score += 10;
reasons.push('reasoning model for complex task');
}
}
// Large context bonus for analysis tasks
if (criteria.taskType === 'analysis' && model.contextWindow >= 128_000) {
score += 5;
reasons.push('large context window');
}
return {
provider: model.provider,
modelId: model.id,
modelName: model.name,
score,
reasoning: reasons.length > 0 ? reasons.join('; ') : 'base score',
};
}
private classifyTier(model: ModelInfo): CostTier {
const cost = model.cost.input;
const cheapThreshold = COST_TIER_THRESHOLDS['cheap'];
const standardThreshold = COST_TIER_THRESHOLDS['standard'];
if (cost <= cheapThreshold.maxInput) return 'cheap';
if (cost <= standardThreshold.maxInput) return 'standard';
return 'premium';
}
private disqualified(model: ModelInfo, reason: string): RoutingResult {
return {
provider: model.provider,
modelId: model.id,
modelName: model.name,
score: 0,
reasoning: `disqualified: ${reason}`,
};
}
}
-44
View File
@@ -1,44 +0,0 @@
export interface SessionInfoDto {
id: string;
provider: string;
modelId: string;
createdAt: string;
promptCount: number;
channels: string[];
durationMs: number;
}
export interface SessionListDto {
sessions: SessionInfoDto[];
total: number;
}
/**
* Options accepted when creating an agent session.
* All fields are optional; omitting them falls back to env-var or process defaults.
*/
export interface CreateSessionOptionsDto {
/** Provider name (e.g. "anthropic", "openai"). */
provider?: string;
/** Model ID to use for this session. */
modelId?: string;
/**
* Sandbox working directory for the session.
* File, git, and shell tools will be restricted to this directory.
* Defaults to AGENT_FILE_SANDBOX_DIR env var or process.cwd().
*/
sandboxDir?: string;
/**
* Platform-level system prompt for this session.
* Merged with skill prompt additions (platform prompt first, then skills).
* Falls back to AGENT_SYSTEM_PROMPT env var when omitted.
*/
systemPrompt?: string;
/**
* Explicit allowlist of tool names available in this session.
* When provided, only listed tools are registered with the agent.
* Admins receive all tools; regular users fall back to AGENT_USER_TOOLS
* env var (comma-separated) when this field is not supplied.
*/
allowedTools?: string[];
}
@@ -1,40 +0,0 @@
import {
Controller,
Delete,
Get,
HttpCode,
HttpStatus,
Inject,
NotFoundException,
Param,
UseGuards,
} from '@nestjs/common';
import { AuthGuard } from '../auth/auth.guard.js';
import { AgentService } from './agent.service.js';
@Controller('api/sessions')
@UseGuards(AuthGuard)
export class SessionsController {
constructor(@Inject(AgentService) private readonly agentService: AgentService) {}
@Get()
list() {
const sessions = this.agentService.listSessions();
return { sessions, total: sessions.length };
}
@Get(':id')
findOne(@Param('id') id: string) {
const info = this.agentService.getSessionInfo(id);
if (!info) throw new NotFoundException('Session not found');
return info;
}
@Delete(':id')
@HttpCode(HttpStatus.NO_CONTENT)
async destroy(@Param('id') id: string) {
const info = this.agentService.getSessionInfo(id);
if (!info) throw new NotFoundException('Session not found');
await this.agentService.destroySession(id);
}
}
@@ -1,59 +0,0 @@
import { Inject, Injectable, Logger } from '@nestjs/common';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import { SkillsService } from '../skills/skills.service.js';
import { createSkillTools } from './tools/skill-tools.js';
export interface LoadedSkills {
/** Meta-tools: skill_list + skill_invoke */
metaTools: ToolDefinition[];
/**
* System prompt additions from enabled prompt-type skills.
* Callers may prepend these to the session system prompt.
*/
promptAdditions: string[];
}
/**
* SkillLoaderService is responsible for:
* 1. Providing the skill meta-tools (skill_list, skill_invoke) to agent sessions.
* 2. Collecting system-prompt additions from enabled prompt-type skills.
*/
@Injectable()
export class SkillLoaderService {
private readonly logger = new Logger(SkillLoaderService.name);
constructor(@Inject(SkillsService) private readonly skillsService: SkillsService) {}
/**
* Load enabled skills and return tools + prompt additions for a new session.
*/
async loadForSession(): Promise<LoadedSkills> {
const metaTools = createSkillTools(this.skillsService);
let promptAdditions: string[] = [];
try {
const enabledSkills = await this.skillsService.findEnabled();
promptAdditions = enabledSkills.flatMap((skill) => {
const config = (skill.config ?? {}) as Record<string, unknown>;
const skillType = (config['type'] as string | undefined) ?? 'prompt';
if (skillType === 'prompt') {
const addition = (config['prompt'] as string | undefined) ?? skill.description;
return addition ? [addition] : [];
}
return [];
});
this.logger.log(
`Loaded ${enabledSkills.length} enabled skill(s), ` +
`${promptAdditions.length} prompt addition(s)`,
);
} catch (err) {
// Non-fatal: log and continue without prompt additions
this.logger.warn(
`Failed to load skill prompt additions: ${err instanceof Error ? err.message : String(err)}`,
);
}
return { metaTools, promptAdditions };
}
}
-184
View File
@@ -1,184 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import type { Brain } from '@mosaic/brain';
export function createBrainTools(brain: Brain): ToolDefinition[] {
const listProjects: ToolDefinition = {
name: 'brain_list_projects',
label: 'List Projects',
description: 'List all projects in the brain.',
parameters: Type.Object({}),
async execute() {
const projects = await brain.projects.findAll();
return {
content: [{ type: 'text' as const, text: JSON.stringify(projects, null, 2) }],
details: undefined,
};
},
};
const getProject: ToolDefinition = {
name: 'brain_get_project',
label: 'Get Project',
description: 'Get a project by ID.',
parameters: Type.Object({
id: Type.String({ description: 'Project ID (UUID)' }),
}),
async execute(_toolCallId, params) {
const { id } = params as { id: string };
const project = await brain.projects.findById(id);
return {
content: [
{
type: 'text' as const,
text: project ? JSON.stringify(project, null, 2) : `Project not found: ${id}`,
},
],
details: undefined,
};
},
};
const listTasks: ToolDefinition = {
name: 'brain_list_tasks',
label: 'List Tasks',
description: 'List tasks, optionally filtered by project, mission, or status.',
parameters: Type.Object({
projectId: Type.Optional(Type.String({ description: 'Filter by project ID' })),
missionId: Type.Optional(Type.String({ description: 'Filter by mission ID' })),
status: Type.Optional(Type.String({ description: 'Filter by status' })),
}),
async execute(_toolCallId, params) {
const p = params as { projectId?: string; missionId?: string; status?: string };
type TaskStatus = 'not-started' | 'in-progress' | 'blocked' | 'done' | 'cancelled';
let tasks;
if (p.projectId) tasks = await brain.tasks.findByProject(p.projectId);
else if (p.missionId) tasks = await brain.tasks.findByMission(p.missionId);
else if (p.status) tasks = await brain.tasks.findByStatus(p.status as TaskStatus);
else tasks = await brain.tasks.findAll();
return {
content: [{ type: 'text' as const, text: JSON.stringify(tasks, null, 2) }],
details: undefined,
};
},
};
const createTask: ToolDefinition = {
name: 'brain_create_task',
label: 'Create Task',
description: 'Create a new task in the brain.',
parameters: Type.Object({
title: Type.String({ description: 'Task title' }),
description: Type.Optional(Type.String({ description: 'Task description' })),
projectId: Type.Optional(Type.String({ description: 'Project ID' })),
missionId: Type.Optional(Type.String({ description: 'Mission ID' })),
priority: Type.Optional(
Type.String({ description: 'Priority: low, medium, high, critical' }),
),
}),
async execute(_toolCallId, params) {
const p = params as {
title: string;
description?: string;
projectId?: string;
missionId?: string;
priority?: string;
};
type Priority = 'low' | 'medium' | 'high' | 'critical';
const task = await brain.tasks.create({
...p,
priority: p.priority as Priority | undefined,
});
return {
content: [{ type: 'text' as const, text: JSON.stringify(task, null, 2) }],
details: undefined,
};
},
};
const updateTask: ToolDefinition = {
name: 'brain_update_task',
label: 'Update Task',
description: 'Update an existing task.',
parameters: Type.Object({
id: Type.String({ description: 'Task ID' }),
title: Type.Optional(Type.String()),
description: Type.Optional(Type.String()),
status: Type.Optional(
Type.String({ description: 'not-started, in-progress, blocked, done, cancelled' }),
),
priority: Type.Optional(Type.String()),
}),
async execute(_toolCallId, params) {
const { id, ...updates } = params as {
id: string;
title?: string;
description?: string;
status?: string;
priority?: string;
};
type TaskStatus = 'not-started' | 'in-progress' | 'blocked' | 'done' | 'cancelled';
type Priority = 'low' | 'medium' | 'high' | 'critical';
const task = await brain.tasks.update(id, {
...updates,
status: updates.status as TaskStatus | undefined,
priority: updates.priority as Priority | undefined,
});
return {
content: [
{
type: 'text' as const,
text: task ? JSON.stringify(task, null, 2) : `Task not found: ${id}`,
},
],
details: undefined,
};
},
};
const listMissions: ToolDefinition = {
name: 'brain_list_missions',
label: 'List Missions',
description: 'List all missions, optionally filtered by project.',
parameters: Type.Object({
projectId: Type.Optional(Type.String({ description: 'Filter by project ID' })),
}),
async execute(_toolCallId, params) {
const p = params as { projectId?: string };
const missions = p.projectId
? await brain.missions.findByProject(p.projectId)
: await brain.missions.findAll();
return {
content: [{ type: 'text' as const, text: JSON.stringify(missions, null, 2) }],
details: undefined,
};
},
};
const listConversations: ToolDefinition = {
name: 'brain_list_conversations',
label: 'List Conversations',
description: 'List conversations for a user.',
parameters: Type.Object({
userId: Type.String({ description: 'User ID' }),
}),
async execute(_toolCallId, params) {
const { userId } = params as { userId: string };
const conversations = await brain.conversations.findAll(userId);
return {
content: [{ type: 'text' as const, text: JSON.stringify(conversations, null, 2) }],
details: undefined,
};
},
};
return [
listProjects,
getProject,
listTasks,
createTask,
updateTask,
listMissions,
listConversations,
];
}
@@ -1,81 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import type { CoordService } from '../../coord/coord.service.js';
export function createCoordTools(coordService: CoordService): ToolDefinition[] {
const getMissionStatus: ToolDefinition = {
name: 'coord_mission_status',
label: 'Mission Status',
description:
'Get the current orchestration mission status including milestones, tasks, and active session.',
parameters: Type.Object({
projectPath: Type.Optional(
Type.String({ description: 'Project path. Defaults to gateway working directory.' }),
),
}),
async execute(_toolCallId, params) {
const { projectPath } = params as { projectPath?: string };
const resolvedPath = projectPath ?? process.cwd();
const status = await coordService.getMissionStatus(resolvedPath);
return {
content: [
{
type: 'text' as const,
text: status ? JSON.stringify(status, null, 2) : 'No active coord mission found.',
},
],
details: undefined,
};
},
};
const listCoordTasks: ToolDefinition = {
name: 'coord_list_tasks',
label: 'List Coord Tasks',
description: 'List all tasks from the orchestration TASKS.md file.',
parameters: Type.Object({
projectPath: Type.Optional(
Type.String({ description: 'Project path. Defaults to gateway working directory.' }),
),
}),
async execute(_toolCallId, params) {
const { projectPath } = params as { projectPath?: string };
const resolvedPath = projectPath ?? process.cwd();
const tasks = await coordService.listTasks(resolvedPath);
return {
content: [{ type: 'text' as const, text: JSON.stringify(tasks, null, 2) }],
details: undefined,
};
},
};
const getCoordTaskDetail: ToolDefinition = {
name: 'coord_task_detail',
label: 'Coord Task Detail',
description: 'Get detailed status for a specific orchestration task.',
parameters: Type.Object({
taskId: Type.String({ description: 'Task ID (e.g. P2-005)' }),
projectPath: Type.Optional(
Type.String({ description: 'Project path. Defaults to gateway working directory.' }),
),
}),
async execute(_toolCallId, params) {
const { taskId, projectPath } = params as { taskId: string; projectPath?: string };
const resolvedPath = projectPath ?? process.cwd();
const detail = await coordService.getTaskStatus(resolvedPath, taskId);
return {
content: [
{
type: 'text' as const,
text: detail
? JSON.stringify(detail, null, 2)
: `Task ${taskId} not found in coord mission.`,
},
],
details: undefined,
};
},
};
return [getMissionStatus, listCoordTasks, getCoordTaskDetail];
}
-194
View File
@@ -1,194 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import { readFile, writeFile, readdir, stat } from 'node:fs/promises';
import { guardPath, guardPathUnsafe, SandboxEscapeError } from './path-guard.js';
const MAX_READ_BYTES = 512 * 1024; // 512 KB read limit
const MAX_WRITE_BYTES = 1024 * 1024; // 1 MB write limit
export function createFileTools(baseDir: string): ToolDefinition[] {
const readFileTool: ToolDefinition = {
name: 'fs_read_file',
label: 'Read File',
description:
'Read the contents of a file. Path is resolved relative to the sandbox base directory.',
parameters: Type.Object({
path: Type.String({
description: 'File path (relative to sandbox base or absolute within it)',
}),
encoding: Type.Optional(
Type.String({ description: 'Encoding: utf8 (default), base64, hex' }),
),
}),
async execute(_toolCallId, params) {
const { path, encoding } = params as { path: string; encoding?: string };
let safePath: string;
try {
safePath = guardPath(path, baseDir);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
try {
const info = await stat(safePath);
if (!info.isFile()) {
return {
content: [{ type: 'text' as const, text: `Error: path is not a file: ${path}` }],
details: undefined,
};
}
if (info.size > MAX_READ_BYTES) {
return {
content: [
{
type: 'text' as const,
text: `Error: file too large (${info.size} bytes, limit ${MAX_READ_BYTES} bytes)`,
},
],
details: undefined,
};
}
const enc = (encoding ?? 'utf8') as BufferEncoding;
const content = await readFile(safePath, { encoding: enc });
return {
content: [{ type: 'text' as const, text: String(content) }],
details: undefined,
};
} catch (err) {
return {
content: [{ type: 'text' as const, text: `Error reading file: ${String(err)}` }],
details: undefined,
};
}
},
};
const writeFileTool: ToolDefinition = {
name: 'fs_write_file',
label: 'Write File',
description:
'Write content to a file. Path is resolved relative to the sandbox base directory. Overwrites existing file.',
parameters: Type.Object({
path: Type.String({
description: 'File path (relative to sandbox base or absolute within it)',
}),
content: Type.String({ description: 'Content to write' }),
encoding: Type.Optional(Type.String({ description: 'Encoding: utf8 (default), base64' })),
}),
async execute(_toolCallId, params) {
const { path, content, encoding } = params as {
path: string;
content: string;
encoding?: string;
};
let safePath: string;
try {
safePath = guardPathUnsafe(path, baseDir);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
if (Buffer.byteLength(content, 'utf8') > MAX_WRITE_BYTES) {
return {
content: [
{
type: 'text' as const,
text: `Error: content too large (limit ${MAX_WRITE_BYTES} bytes)`,
},
],
details: undefined,
};
}
try {
const enc = (encoding ?? 'utf8') as BufferEncoding;
await writeFile(safePath, content, { encoding: enc });
return {
content: [{ type: 'text' as const, text: `File written successfully: ${path}` }],
details: undefined,
};
} catch (err) {
return {
content: [{ type: 'text' as const, text: `Error writing file: ${String(err)}` }],
details: undefined,
};
}
},
};
const listDirectoryTool: ToolDefinition = {
name: 'fs_list_directory',
label: 'List Directory',
description: 'List files and directories at a given path within the sandbox base directory.',
parameters: Type.Object({
path: Type.Optional(
Type.String({
description: 'Directory path (relative to sandbox base). Defaults to base directory.',
}),
),
}),
async execute(_toolCallId, params) {
const { path } = params as { path?: string };
const target = path ?? '.';
let safePath: string;
try {
safePath = guardPath(target, baseDir);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
try {
const info = await stat(safePath);
if (!info.isDirectory()) {
return {
content: [{ type: 'text' as const, text: `Error: path is not a directory: ${target}` }],
details: undefined,
};
}
const entries = await readdir(safePath, { withFileTypes: true });
const items = entries.map((e) => ({
name: e.name,
type: e.isDirectory() ? 'directory' : e.isSymbolicLink() ? 'symlink' : 'file',
}));
return {
content: [{ type: 'text' as const, text: JSON.stringify(items, null, 2) }],
details: undefined,
};
} catch (err) {
return {
content: [{ type: 'text' as const, text: `Error listing directory: ${String(err)}` }],
details: undefined,
};
}
},
};
return [readFileTool, writeFileTool, listDirectoryTool];
}
-212
View File
@@ -1,212 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import { exec } from 'node:child_process';
import { promisify } from 'node:util';
import { guardPath, guardPathUnsafe, SandboxEscapeError } from './path-guard.js';
const execAsync = promisify(exec);
const GIT_TIMEOUT_MS = 15_000;
const MAX_OUTPUT_BYTES = 100 * 1024; // 100 KB
async function runGit(
args: string[],
cwd?: string,
): Promise<{ stdout: string; stderr: string; error?: string }> {
// Only allow specific safe read-only git subcommands
const allowedSubcommands = ['status', 'log', 'diff', 'show', 'branch', 'tag', 'ls-files'];
const subcommand = args[0];
if (!subcommand || !allowedSubcommands.includes(subcommand)) {
return {
stdout: '',
stderr: '',
error: `Blocked: git subcommand "${subcommand}" is not allowed. Permitted: ${allowedSubcommands.join(', ')}`,
};
}
const cmd = `git ${args.map((a) => JSON.stringify(a)).join(' ')}`;
try {
const { stdout, stderr } = await execAsync(cmd, {
cwd,
timeout: GIT_TIMEOUT_MS,
maxBuffer: MAX_OUTPUT_BYTES,
});
return { stdout, stderr };
} catch (err: unknown) {
const e = err as { stdout?: string; stderr?: string; message?: string };
return {
stdout: e.stdout ?? '',
stderr: e.stderr ?? '',
error: e.message ?? String(err),
};
}
}
export function createGitTools(sandboxDir?: string): ToolDefinition[] {
const defaultCwd = sandboxDir ?? process.cwd();
const gitStatus: ToolDefinition = {
name: 'git_status',
label: 'Git Status',
description: 'Show the working tree status (staged, unstaged, untracked files).',
parameters: Type.Object({
cwd: Type.Optional(
Type.String({
description: 'Repository working directory (relative to sandbox or absolute within it).',
}),
),
}),
async execute(_toolCallId, params) {
const { cwd } = params as { cwd?: string };
let safeCwd: string;
try {
safeCwd = guardPath(cwd ?? '.', defaultCwd);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
const result = await runGit(['status', '--short', '--branch'], safeCwd);
const text = result.error
? `Error: ${result.error}\n${result.stderr}`
: result.stdout || '(no output)';
return {
content: [{ type: 'text' as const, text: text }],
details: undefined,
};
},
};
const gitLog: ToolDefinition = {
name: 'git_log',
label: 'Git Log',
description: 'Show recent commit history.',
parameters: Type.Object({
limit: Type.Optional(Type.Number({ description: 'Number of commits to show (default 20)' })),
oneline: Type.Optional(
Type.Boolean({ description: 'Compact one-line format (default true)' }),
),
cwd: Type.Optional(
Type.String({
description: 'Repository working directory (relative to sandbox or absolute within it).',
}),
),
}),
async execute(_toolCallId, params) {
const { limit, oneline, cwd } = params as {
limit?: number;
oneline?: boolean;
cwd?: string;
};
let safeCwd: string;
try {
safeCwd = guardPath(cwd ?? '.', defaultCwd);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
const args = ['log', `--max-count=${limit ?? 20}`];
if (oneline !== false) args.push('--oneline');
const result = await runGit(args, safeCwd);
const text = result.error
? `Error: ${result.error}\n${result.stderr}`
: result.stdout || '(no commits)';
return {
content: [{ type: 'text' as const, text: text }],
details: undefined,
};
},
};
const gitDiff: ToolDefinition = {
name: 'git_diff',
label: 'Git Diff',
description: 'Show changes between commits, working tree, or staged changes.',
parameters: Type.Object({
staged: Type.Optional(
Type.Boolean({ description: 'Show staged (cached) changes instead of unstaged' }),
),
ref: Type.Optional(
Type.String({ description: 'Compare against this ref (commit SHA, branch, or tag)' }),
),
path: Type.Optional(
Type.String({ description: 'Limit diff to a specific file or directory' }),
),
cwd: Type.Optional(
Type.String({
description: 'Repository working directory (relative to sandbox or absolute within it).',
}),
),
}),
async execute(_toolCallId, params) {
const { staged, ref, path, cwd } = params as {
staged?: boolean;
ref?: string;
path?: string;
cwd?: string;
};
let safeCwd: string;
try {
safeCwd = guardPath(cwd ?? '.', defaultCwd);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
let safePath: string | undefined;
if (path !== undefined) {
try {
safePath = guardPathUnsafe(path, defaultCwd);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
}
const args = ['diff'];
if (staged) args.push('--cached');
if (ref) args.push(ref);
args.push('--');
if (safePath !== undefined) args.push(safePath);
const result = await runGit(args, safeCwd);
const text = result.error
? `Error: ${result.error}\n${result.stderr}`
: result.stdout || '(no diff)';
return {
content: [{ type: 'text' as const, text: text }],
details: undefined,
};
},
};
return [gitStatus, gitLog, gitDiff];
}
-7
View File
@@ -1,7 +0,0 @@
export { createBrainTools } from './brain-tools.js';
export { createCoordTools } from './coord-tools.js';
export { createFileTools } from './file-tools.js';
export { createGitTools } from './git-tools.js';
export { createShellTools } from './shell-tools.js';
export { createWebTools } from './web-tools.js';
export { createSkillTools } from './skill-tools.js';
@@ -1,158 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import type { Memory } from '@mosaic/memory';
import type { EmbeddingProvider } from '@mosaic/memory';
export function createMemoryTools(
memory: Memory,
embeddingProvider: EmbeddingProvider | null,
): ToolDefinition[] {
const searchMemory: ToolDefinition = {
name: 'memory_search',
label: 'Search Memory',
description:
'Search across stored insights and knowledge using natural language. Returns semantically similar results.',
parameters: Type.Object({
userId: Type.String({ description: 'User ID to search memory for' }),
query: Type.String({ description: 'Natural language search query' }),
limit: Type.Optional(Type.Number({ description: 'Max results (default 5)' })),
}),
async execute(_toolCallId, params) {
const { userId, query, limit } = params as {
userId: string;
query: string;
limit?: number;
};
if (!embeddingProvider) {
return {
content: [
{
type: 'text' as const,
text: 'Semantic search unavailable — no embedding provider configured',
},
],
details: undefined,
};
}
const embedding = await embeddingProvider.embed(query);
const results = await memory.insights.searchByEmbedding(userId, embedding, limit ?? 5);
return {
content: [{ type: 'text' as const, text: JSON.stringify(results, null, 2) }],
details: undefined,
};
},
};
const getPreferences: ToolDefinition = {
name: 'memory_get_preferences',
label: 'Get User Preferences',
description: 'Retrieve stored preferences for a user.',
parameters: Type.Object({
userId: Type.String({ description: 'User ID' }),
category: Type.Optional(
Type.String({
description: 'Filter by category: communication, coding, workflow, appearance, general',
}),
),
}),
async execute(_toolCallId, params) {
const { userId, category } = params as { userId: string; category?: string };
type Cat = 'communication' | 'coding' | 'workflow' | 'appearance' | 'general';
const prefs = category
? await memory.preferences.findByUserAndCategory(userId, category as Cat)
: await memory.preferences.findByUser(userId);
return {
content: [{ type: 'text' as const, text: JSON.stringify(prefs, null, 2) }],
details: undefined,
};
},
};
const savePreference: ToolDefinition = {
name: 'memory_save_preference',
label: 'Save User Preference',
description:
'Store a learned user preference (e.g., "prefers tables over paragraphs", "timezone: America/Chicago").',
parameters: Type.Object({
userId: Type.String({ description: 'User ID' }),
key: Type.String({ description: 'Preference key' }),
value: Type.String({ description: 'Preference value (JSON string)' }),
category: Type.Optional(
Type.String({
description: 'Category: communication, coding, workflow, appearance, general',
}),
),
}),
async execute(_toolCallId, params) {
const { userId, key, value, category } = params as {
userId: string;
key: string;
value: string;
category?: string;
};
type Cat = 'communication' | 'coding' | 'workflow' | 'appearance' | 'general';
let parsedValue: unknown;
try {
parsedValue = JSON.parse(value);
} catch {
parsedValue = value;
}
const pref = await memory.preferences.upsert({
userId,
key,
value: parsedValue,
category: (category as Cat) ?? 'general',
source: 'agent',
});
return {
content: [{ type: 'text' as const, text: JSON.stringify(pref, null, 2) }],
details: undefined,
};
},
};
const saveInsight: ToolDefinition = {
name: 'memory_save_insight',
label: 'Save Insight',
description:
'Store a learned insight, decision, or knowledge extracted from the current interaction.',
parameters: Type.Object({
userId: Type.String({ description: 'User ID' }),
content: Type.String({ description: 'The insight or knowledge to store' }),
category: Type.Optional(
Type.String({
description: 'Category: decision, learning, preference, fact, pattern, general',
}),
),
}),
async execute(_toolCallId, params) {
const { userId, content, category } = params as {
userId: string;
content: string;
category?: string;
};
type Cat = 'decision' | 'learning' | 'preference' | 'fact' | 'pattern' | 'general';
let embedding: number[] | null = null;
if (embeddingProvider) {
embedding = await embeddingProvider.embed(content);
}
const insight = await memory.insights.create({
userId,
content,
embedding,
source: 'agent',
category: (category as Cat) ?? 'learning',
});
return {
content: [{ type: 'text' as const, text: JSON.stringify(insight, null, 2) }],
details: undefined,
};
},
};
return [searchMemory, getPreferences, savePreference, saveInsight];
}
@@ -1,104 +0,0 @@
import { describe, it, expect } from 'vitest';
import { guardPath, guardPathUnsafe, SandboxEscapeError } from './path-guard.js';
import path from 'node:path';
import os from 'node:os';
import fs from 'node:fs';
describe('guardPathUnsafe', () => {
const sandbox = '/tmp/test-sandbox';
it('allows paths inside sandbox', () => {
const result = guardPathUnsafe('foo/bar.txt', sandbox);
expect(result).toBe(path.resolve(sandbox, 'foo/bar.txt'));
});
it('allows sandbox root itself', () => {
const result = guardPathUnsafe('.', sandbox);
expect(result).toBe(path.resolve(sandbox));
});
it('rejects path traversal with ../', () => {
expect(() => guardPathUnsafe('../escape.txt', sandbox)).toThrow(SandboxEscapeError);
});
it('rejects absolute path outside sandbox', () => {
expect(() => guardPathUnsafe('/etc/passwd', sandbox)).toThrow(SandboxEscapeError);
});
it('rejects deeply nested traversal', () => {
expect(() => guardPathUnsafe('a/b/../../../../../../etc/passwd', sandbox)).toThrow(
SandboxEscapeError,
);
});
it('rejects path that starts with sandbox name but is sibling', () => {
expect(() => guardPathUnsafe('/tmp/test-sandbox-evil/file.txt', sandbox)).toThrow(
SandboxEscapeError,
);
});
it('returns the resolved absolute path for nested paths', () => {
const result = guardPathUnsafe('deep/nested/file.ts', sandbox);
expect(result).toBe('/tmp/test-sandbox/deep/nested/file.ts');
});
it('SandboxEscapeError includes the user path and sandbox in message', () => {
let caught: unknown;
try {
guardPathUnsafe('../escape.txt', sandbox);
} catch (err) {
caught = err;
}
expect(caught).toBeInstanceOf(SandboxEscapeError);
const e = caught as SandboxEscapeError;
expect(e.userPath).toBe('../escape.txt');
expect(e.sandboxDir).toBe(sandbox);
expect(e.message).toContain('Path escape attempt blocked');
});
});
describe('guardPath', () => {
let tmpDir: string;
it('allows an existing path inside a real temp sandbox', () => {
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'path-guard-test-'));
try {
const subdir = path.join(tmpDir, 'subdir');
fs.mkdirSync(subdir);
const result = guardPath('subdir', tmpDir);
expect(result).toBe(subdir);
} finally {
fs.rmSync(tmpDir, { recursive: true, force: true });
}
});
it('allows sandbox root itself', () => {
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'path-guard-test-'));
try {
const result = guardPath('.', tmpDir);
// realpathSync resolves the tmpdir symlinks (macOS /var -> /private/var)
const realTmp = fs.realpathSync.native(tmpDir);
expect(result).toBe(realTmp);
} finally {
fs.rmSync(tmpDir, { recursive: true, force: true });
}
});
it('rejects path traversal with ../ on existing sandbox', () => {
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'path-guard-test-'));
try {
expect(() => guardPath('../escape', tmpDir)).toThrow(SandboxEscapeError);
} finally {
fs.rmSync(tmpDir, { recursive: true, force: true });
}
});
it('rejects absolute path outside sandbox', () => {
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'path-guard-test-'));
try {
expect(() => guardPath('/etc/passwd', tmpDir)).toThrow(SandboxEscapeError);
} finally {
fs.rmSync(tmpDir, { recursive: true, force: true });
}
});
});
@@ -1,58 +0,0 @@
import path from 'node:path';
import fs from 'node:fs';
/**
* Resolves a user-provided path and verifies it is inside the allowed sandbox directory.
* Throws SandboxEscapeError if the resolved path is outside the sandbox.
*
* Uses realpathSync to resolve symlinks in the sandbox root. The user-supplied path
* is checked for containment AFTER lexical resolution but BEFORE resolving any symlinks
* within the user path — so symlink escape attempts are caught too.
*
* @param userPath - The path provided by the agent (may be relative or absolute)
* @param sandboxDir - The allowed root directory (already validated on session creation)
* @returns The resolved absolute path, guaranteed to be within sandboxDir
*/
export function guardPath(userPath: string, sandboxDir: string): string {
const resolved = path.resolve(sandboxDir, userPath);
const sandboxResolved = fs.realpathSync.native(sandboxDir);
// Normalize both paths to resolve any symlinks in the sandbox root itself.
// For the user path, we check containment BEFORE resolving symlinks in the path
// (so we catch symlink escape attempts too — the resolved path must still be under sandbox)
if (!resolved.startsWith(sandboxResolved + path.sep) && resolved !== sandboxResolved) {
throw new SandboxEscapeError(userPath, sandboxDir, resolved);
}
return resolved;
}
/**
* Validates a path without resolving symlinks in the user-provided portion.
* Use for paths that may not exist yet (creates, writes).
*
* Performs a lexical containment check only using path.resolve.
*/
export function guardPathUnsafe(userPath: string, sandboxDir: string): string {
const resolved = path.resolve(sandboxDir, userPath);
const sandboxAbs = path.resolve(sandboxDir);
if (!resolved.startsWith(sandboxAbs + path.sep) && resolved !== sandboxAbs) {
throw new SandboxEscapeError(userPath, sandboxDir, resolved);
}
return resolved;
}
export class SandboxEscapeError extends Error {
constructor(
public readonly userPath: string,
public readonly sandboxDir: string,
public readonly resolvedPath: string,
) {
super(
`Path escape attempt blocked: "${userPath}" resolves to "${resolvedPath}" which is outside sandbox "${sandboxDir}"`,
);
this.name = 'SandboxEscapeError';
}
}
-218
View File
@@ -1,218 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import { spawn } from 'node:child_process';
import { guardPath, SandboxEscapeError } from './path-guard.js';
const DEFAULT_TIMEOUT_MS = 30_000;
const MAX_OUTPUT_BYTES = 100 * 1024; // 100 KB
/**
* Commands that are outright blocked for safety.
* This is a denylist; the agent should be instructed to use
* the least-privilege command necessary.
*/
const BLOCKED_COMMANDS = new Set([
'rm',
'rmdir',
'mkfs',
'dd',
'format',
'fdisk',
'parted',
'shred',
'wipefs',
'sudo',
'su',
'chown',
'chmod',
'passwd',
'useradd',
'userdel',
'groupadd',
'shutdown',
'reboot',
'halt',
'poweroff',
'kill',
'killall',
'pkill',
'curl',
'wget',
'nc',
'netcat',
'ncat',
'ssh',
'scp',
'sftp',
'rsync',
'iptables',
'ip6tables',
'nft',
'ufw',
'firewall-cmd',
'docker',
'podman',
'kubectl',
'helm',
'terraform',
'ansible',
'crontab',
'at',
'batch',
]);
function extractBaseCommand(command: string): string {
// Extract the first word (the binary name), stripping path
const trimmed = command.trim();
const firstToken = trimmed.split(/\s+/)[0] ?? '';
return firstToken.split('/').pop() ?? firstToken;
}
function runCommand(
command: string,
options: { timeoutMs: number; cwd?: string },
): Promise<{ stdout: string; stderr: string; exitCode: number | null; timedOut: boolean }> {
return new Promise((resolve) => {
const child = spawn('sh', ['-c', command], {
cwd: options.cwd,
stdio: ['ignore', 'pipe', 'pipe'],
detached: false,
});
let stdout = '';
let stderr = '';
let timedOut = false;
let totalBytes = 0;
let truncated = false;
child.stdout?.on('data', (chunk: Buffer) => {
if (truncated) return;
totalBytes += chunk.length;
if (totalBytes > MAX_OUTPUT_BYTES) {
stdout += chunk.subarray(0, MAX_OUTPUT_BYTES - (totalBytes - chunk.length)).toString();
stdout += '\n[output truncated at 100 KB limit]';
truncated = true;
child.kill('SIGTERM');
} else {
stdout += chunk.toString();
}
});
child.stderr?.on('data', (chunk: Buffer) => {
if (stderr.length < MAX_OUTPUT_BYTES) {
stderr += chunk.toString();
}
});
const timer = setTimeout(() => {
timedOut = true;
child.kill('SIGTERM');
setTimeout(() => {
try {
child.kill('SIGKILL');
} catch {
// already exited
}
}, 2000);
}, options.timeoutMs);
child.on('close', (exitCode) => {
clearTimeout(timer);
resolve({ stdout, stderr, exitCode, timedOut });
});
child.on('error', (err) => {
clearTimeout(timer);
resolve({ stdout, stderr: stderr + String(err), exitCode: null, timedOut: false });
});
});
}
export function createShellTools(sandboxDir?: string): ToolDefinition[] {
const defaultCwd = sandboxDir ?? process.cwd();
const shellExec: ToolDefinition = {
name: 'shell_exec',
label: 'Shell Execute',
description:
'Execute a shell command with timeout and output limits. Dangerous commands (rm, sudo, docker, etc.) are blocked. Working directory is restricted to the session sandbox.',
parameters: Type.Object({
command: Type.String({ description: 'Shell command to execute' }),
cwd: Type.Optional(
Type.String({
description:
'Working directory for the command (relative to sandbox or absolute within it).',
}),
),
timeout: Type.Optional(
Type.Number({ description: 'Timeout in milliseconds (default 30000, max 60000)' }),
),
}),
async execute(_toolCallId, params) {
const { command, cwd, timeout } = params as {
command: string;
cwd?: string;
timeout?: number;
};
const base = extractBaseCommand(command);
if (BLOCKED_COMMANDS.has(base)) {
return {
content: [
{
type: 'text' as const,
text: `Error: command "${base}" is blocked for safety reasons.`,
},
],
details: undefined,
};
}
const timeoutMs = Math.min(timeout ?? DEFAULT_TIMEOUT_MS, 60_000);
let safeCwd: string;
try {
safeCwd = guardPath(cwd ?? '.', defaultCwd);
} catch (err) {
if (err instanceof SandboxEscapeError) {
return {
content: [{ type: 'text' as const, text: `Error: ${err.message}` }],
details: undefined,
};
}
return {
content: [{ type: 'text' as const, text: `Error: ${String(err)}` }],
details: undefined,
};
}
const result = await runCommand(command, {
timeoutMs,
cwd: safeCwd,
});
if (result.timedOut) {
return {
content: [
{
type: 'text' as const,
text: `Command timed out after ${timeoutMs}ms.\nPartial stdout:\n${result.stdout}\nPartial stderr:\n${result.stderr}`,
},
],
details: undefined,
};
}
const parts: string[] = [];
if (result.stdout) parts.push(`stdout:\n${result.stdout}`);
if (result.stderr) parts.push(`stderr:\n${result.stderr}`);
parts.push(`exit code: ${result.exitCode ?? 'null'}`);
return {
content: [{ type: 'text' as const, text: parts.join('\n') }],
details: undefined,
};
},
};
return [shellExec];
}
-180
View File
@@ -1,180 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
import type { SkillsService } from '../../skills/skills.service.js';
/**
* Creates meta-tools that allow agents to list and invoke skills from the catalog.
*
* skill_list — list all enabled skills
* skill_invoke — execute a skill by name with parameters
*/
export function createSkillTools(skillsService: SkillsService): ToolDefinition[] {
const skillList: ToolDefinition = {
name: 'skill_list',
label: 'List Skills',
description:
'List all enabled skills available in the catalog. Returns name, description, type, and config for each skill.',
parameters: Type.Object({}),
async execute() {
const skills = await skillsService.findEnabled();
const summary = skills.map((s) => ({
name: s.name,
description: s.description,
version: s.version,
source: s.source,
config: s.config,
}));
return {
content: [
{
type: 'text' as const,
text:
summary.length > 0
? JSON.stringify(summary, null, 2)
: 'No enabled skills found in catalog.',
},
],
details: undefined,
};
},
};
const skillInvoke: ToolDefinition = {
name: 'skill_invoke',
label: 'Invoke Skill',
description:
'Invoke a skill from the catalog by name. For prompt skills, returns the prompt addition. ' +
'For tool skills, executes the embedded logic. For workflow skills, returns the workflow steps.',
parameters: Type.Object({
name: Type.String({ description: 'Skill name to invoke' }),
params: Type.Optional(
Type.Record(Type.String(), Type.Unknown(), {
description: 'Parameters to pass to the skill (if applicable)',
}),
),
}),
async execute(_toolCallId, rawParams) {
const { name, params } = rawParams as {
name: string;
params?: Record<string, unknown>;
};
const skill = await skillsService.findByName(name);
if (!skill) {
return {
content: [{ type: 'text' as const, text: `Skill not found: ${name}` }],
details: undefined,
};
}
if (!skill.enabled) {
return {
content: [{ type: 'text' as const, text: `Skill is disabled: ${name}` }],
details: undefined,
};
}
const config = (skill.config ?? {}) as Record<string, unknown>;
const skillType = (config['type'] as string | undefined) ?? 'prompt';
switch (skillType) {
case 'prompt': {
const promptAddition =
(config['prompt'] as string | undefined) ?? skill.description ?? '';
return {
content: [
{
type: 'text' as const,
text: promptAddition
? `[Skill: ${name}] ${promptAddition}`
: `[Skill: ${name}] No prompt content defined.`,
},
],
details: undefined,
};
}
case 'tool': {
const toolLogic = config['logic'] as string | undefined;
if (!toolLogic) {
return {
content: [
{
type: 'text' as const,
text: `[Skill: ${name}] Tool skill has no logic defined.`,
},
],
details: undefined,
};
}
// Inline tool skill execution: the logic field holds a JS expression or template
// For safety, treat it as a template that can reference params
const result = renderTemplate(toolLogic, { params: params ?? {}, skill });
return {
content: [{ type: 'text' as const, text: `[Skill: ${name}]\n${result}` }],
details: undefined,
};
}
case 'workflow': {
const steps = config['steps'] as unknown[] | undefined;
if (!steps || steps.length === 0) {
return {
content: [
{
type: 'text' as const,
text: `[Skill: ${name}] Workflow has no steps defined.`,
},
],
details: undefined,
};
}
return {
content: [
{
type: 'text' as const,
text: `[Skill: ${name}] Workflow steps:\n${JSON.stringify(steps, null, 2)}`,
},
],
details: undefined,
};
}
default: {
// Unknown type — return full config so the agent can decide what to do
return {
content: [
{
type: 'text' as const,
text: `[Skill: ${name}] (type: ${skillType})\n${JSON.stringify(config, null, 2)}`,
},
],
details: undefined,
};
}
}
},
};
return [skillList, skillInvoke];
}
/**
* Minimal template renderer — replaces {{key}} with values from the context.
* Used for tool skill logic templates.
*/
function renderTemplate(template: string, context: Record<string, unknown>): string {
return template.replace(/\{\{(\w+(?:\.\w+)*)\}\}/g, (_match, path: string) => {
const parts = path.split('.');
let value: unknown = context;
for (const part of parts) {
if (value != null && typeof value === 'object') {
value = (value as Record<string, unknown>)[part];
} else {
value = undefined;
break;
}
}
return value !== undefined && value !== null ? String(value) : '';
});
}
-225
View File
@@ -1,225 +0,0 @@
import { Type } from '@sinclair/typebox';
import type { ToolDefinition } from '@mariozechner/pi-coding-agent';
const DEFAULT_TIMEOUT_MS = 15_000;
const MAX_RESPONSE_BYTES = 512 * 1024; // 512 KB
/**
* Blocked URL patterns (private IP ranges, localhost, link-local).
*/
const BLOCKED_HOSTNAMES = [
/^localhost$/i,
/^127\./,
/^10\./,
/^172\.(1[6-9]|2\d|3[01])\./,
/^192\.168\./,
/^::1$/,
/^fc[0-9a-f][0-9a-f]:/i,
/^fe80:/i,
/^0\.0\.0\.0$/,
/^169\.254\./,
];
function isBlockedUrl(urlString: string): string | null {
let parsed: URL;
try {
parsed = new URL(urlString);
} catch {
return `Invalid URL: ${urlString}`;
}
if (parsed.protocol !== 'http:' && parsed.protocol !== 'https:') {
return `Unsupported protocol: ${parsed.protocol}. Only http and https are allowed.`;
}
const hostname = parsed.hostname;
for (const pattern of BLOCKED_HOSTNAMES) {
if (pattern.test(hostname)) {
return `Blocked: requests to "${hostname}" are not allowed (private/local addresses).`;
}
}
return null;
}
async function fetchWithLimit(
url: string,
options: RequestInit,
timeoutMs: number,
): Promise<{ text: string; status: number; contentType: string }> {
const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), timeoutMs);
try {
const response = await fetch(url, { ...options, signal: controller.signal });
const contentType = response.headers.get('content-type') ?? '';
// Stream response and enforce size limit
const reader = response.body?.getReader();
if (!reader) {
return { text: '', status: response.status, contentType };
}
const chunks: Uint8Array[] = [];
let totalBytes = 0;
let truncated = false;
while (true) {
const { done, value } = await reader.read();
if (done) break;
totalBytes += value.length;
if (totalBytes > MAX_RESPONSE_BYTES) {
const remaining = MAX_RESPONSE_BYTES - (totalBytes - value.length);
chunks.push(value.subarray(0, remaining));
truncated = true;
reader.cancel();
break;
}
chunks.push(value);
}
const combined = new Uint8Array(chunks.reduce((acc, c) => acc + c.length, 0));
let offset = 0;
for (const chunk of chunks) {
combined.set(chunk, offset);
offset += chunk.length;
}
let text = new TextDecoder().decode(combined);
if (truncated) {
text += '\n[response truncated at 512 KB limit]';
}
return { text, status: response.status, contentType };
} finally {
clearTimeout(timer);
}
}
export function createWebTools(): ToolDefinition[] {
const webGet: ToolDefinition = {
name: 'web_get',
label: 'HTTP GET',
description:
'Perform an HTTP GET request and return the response body. Private/local addresses are blocked.',
parameters: Type.Object({
url: Type.String({ description: 'URL to fetch (http/https only)' }),
headers: Type.Optional(
Type.Record(Type.String(), Type.String(), {
description: 'Optional request headers as key-value pairs',
}),
),
timeout: Type.Optional(
Type.Number({ description: 'Timeout in milliseconds (default 15000, max 30000)' }),
),
}),
async execute(_toolCallId, params) {
const { url, headers, timeout } = params as {
url: string;
headers?: Record<string, string>;
timeout?: number;
};
const blocked = isBlockedUrl(url);
if (blocked) {
return {
content: [{ type: 'text' as const, text: `Error: ${blocked}` }],
details: undefined,
};
}
const timeoutMs = Math.min(timeout ?? DEFAULT_TIMEOUT_MS, 30_000);
try {
const result = await fetchWithLimit(
url,
{ method: 'GET', headers: headers ?? {} },
timeoutMs,
);
return {
content: [
{
type: 'text' as const,
text: `HTTP ${result.status} (${result.contentType})\n\n${result.text}`,
},
],
details: undefined,
};
} catch (err) {
const msg = err instanceof Error ? err.message : String(err);
return {
content: [{ type: 'text' as const, text: `Error fetching URL: ${msg}` }],
details: undefined,
};
}
},
};
const webPost: ToolDefinition = {
name: 'web_post',
label: 'HTTP POST',
description:
'Perform an HTTP POST request with a JSON or text body. Private/local addresses are blocked.',
parameters: Type.Object({
url: Type.String({ description: 'URL to POST to (http/https only)' }),
body: Type.String({ description: 'Request body (JSON string or plain text)' }),
contentType: Type.Optional(
Type.String({ description: 'Content-Type header (default: application/json)' }),
),
headers: Type.Optional(
Type.Record(Type.String(), Type.String(), {
description: 'Optional additional request headers',
}),
),
timeout: Type.Optional(
Type.Number({ description: 'Timeout in milliseconds (default 15000, max 30000)' }),
),
}),
async execute(_toolCallId, params) {
const { url, body, contentType, headers, timeout } = params as {
url: string;
body: string;
contentType?: string;
headers?: Record<string, string>;
timeout?: number;
};
const blocked = isBlockedUrl(url);
if (blocked) {
return {
content: [{ type: 'text' as const, text: `Error: ${blocked}` }],
details: undefined,
};
}
const timeoutMs = Math.min(timeout ?? DEFAULT_TIMEOUT_MS, 30_000);
const ct = contentType ?? 'application/json';
try {
const result = await fetchWithLimit(
url,
{
method: 'POST',
headers: { 'Content-Type': ct, ...(headers ?? {}) },
body,
},
timeoutMs,
);
return {
content: [
{
type: 'text' as const,
text: `HTTP ${result.status} (${result.contentType})\n\n${result.text}`,
},
],
details: undefined,
};
} catch (err) {
const msg = err instanceof Error ? err.message : String(err);
return {
content: [{ type: 'text' as const, text: `Error posting to URL: ${msg}` }],
details: undefined,
};
}
},
};
return [webGet, webPost];
}
-60
View File
@@ -1,60 +0,0 @@
import { Module } from '@nestjs/common';
import { APP_GUARD } from '@nestjs/core';
import { HealthController } from './health/health.controller.js';
import { DatabaseModule } from './database/database.module.js';
import { AuthModule } from './auth/auth.module.js';
import { BrainModule } from './brain/brain.module.js';
import { AgentModule } from './agent/agent.module.js';
import { ChatModule } from './chat/chat.module.js';
import { ConversationsModule } from './conversations/conversations.module.js';
import { ProjectsModule } from './projects/projects.module.js';
import { MissionsModule } from './missions/missions.module.js';
import { TasksModule } from './tasks/tasks.module.js';
import { CoordModule } from './coord/coord.module.js';
import { MemoryModule } from './memory/memory.module.js';
import { LogModule } from './log/log.module.js';
import { SkillsModule } from './skills/skills.module.js';
import { PluginModule } from './plugin/plugin.module.js';
import { McpModule } from './mcp/mcp.module.js';
import { AdminModule } from './admin/admin.module.js';
import { CommandsModule } from './commands/commands.module.js';
import { PreferencesModule } from './preferences/preferences.module.js';
import { GCModule } from './gc/gc.module.js';
import { ReloadModule } from './reload/reload.module.js';
import { WorkspaceModule } from './workspace/workspace.module.js';
import { ThrottlerGuard, ThrottlerModule } from '@nestjs/throttler';
@Module({
imports: [
ThrottlerModule.forRoot([{ name: 'default', ttl: 60_000, limit: 60 }]),
DatabaseModule,
AuthModule,
BrainModule,
AgentModule,
ChatModule,
ConversationsModule,
ProjectsModule,
MissionsModule,
TasksModule,
CoordModule,
MemoryModule,
LogModule,
SkillsModule,
PluginModule,
McpModule,
AdminModule,
PreferencesModule,
CommandsModule,
GCModule,
ReloadModule,
WorkspaceModule,
],
controllers: [HealthController],
providers: [
{
provide: APP_GUARD,
useClass: ThrottlerGuard,
},
],
})
export class AppModule {}
-68
View File
@@ -1,68 +0,0 @@
import type { IncomingMessage, ServerResponse } from 'node:http';
import { toNodeHandler } from 'better-auth/node';
import type { Auth } from '@mosaic/auth';
import type { NestFastifyApplication } from '@nestjs/platform-fastify';
import { AUTH } from './auth.tokens.js';
export function mountAuthHandler(app: NestFastifyApplication): void {
const auth = app.get<Auth>(AUTH);
const nodeHandler = toNodeHandler(auth);
const corsOrigin = process.env['GATEWAY_CORS_ORIGIN'] ?? 'http://localhost:3000';
const fastify = app.getHttpAdapter().getInstance();
// BetterAuth is mounted at the raw HTTP level via Fastify's onRequest hook,
// bypassing NestJS middleware (including CORS). We must set CORS headers
// manually on the raw response before handing off to BetterAuth.
fastify.addHook(
'onRequest',
(
req: { raw: IncomingMessage; url: string; method: string },
reply: { raw: ServerResponse; hijack: () => void },
done: () => void,
) => {
if (!req.url.startsWith('/api/auth/')) {
done();
return;
}
const origin = req.raw.headers.origin;
const allowed = corsOrigin.split(',').map((o) => o.trim());
if (origin && allowed.includes(origin)) {
reply.raw.setHeader('Access-Control-Allow-Origin', origin);
reply.raw.setHeader('Access-Control-Allow-Credentials', 'true');
reply.raw.setHeader(
'Access-Control-Allow-Methods',
'GET, POST, PUT, PATCH, DELETE, OPTIONS',
);
reply.raw.setHeader('Access-Control-Allow-Headers', 'Content-Type, Authorization, Cookie');
}
// Handle preflight
if (req.method === 'OPTIONS') {
reply.hijack();
reply.raw.writeHead(204);
reply.raw.end();
return;
}
reply.hijack();
nodeHandler(req.raw as IncomingMessage, reply.raw as ServerResponse)
.then(() => {
if (!reply.raw.writableEnded) {
reply.raw.end();
}
})
.catch((err: unknown) => {
if (!reply.raw.headersSent) {
reply.raw.writeHead(500, { 'Content-Type': 'application/json' });
}
if (!reply.raw.writableEnded) {
reply.raw.end(JSON.stringify({ error: 'Internal auth error' }));
}
console.error('[AUTH] Handler error:', err);
});
},
);
}
-32
View File
@@ -1,32 +0,0 @@
import {
CanActivate,
ExecutionContext,
Inject,
Injectable,
UnauthorizedException,
} from '@nestjs/common';
import { fromNodeHeaders } from 'better-auth/node';
import type { Auth } from '@mosaic/auth';
import type { FastifyRequest } from 'fastify';
import { AUTH } from './auth.tokens.js';
@Injectable()
export class AuthGuard implements CanActivate {
constructor(@Inject(AUTH) private readonly auth: Auth) {}
async canActivate(context: ExecutionContext): Promise<boolean> {
const request = context.switchToHttp().getRequest<FastifyRequest>();
const headers = fromNodeHeaders(request.raw.headers);
const result = await this.auth.api.getSession({ headers });
if (!result) {
throw new UnauthorizedException('Invalid or expired session');
}
(request as FastifyRequest & { user: unknown; session: unknown }).user = result.user;
(request as FastifyRequest & { user: unknown; session: unknown }).session = result.session;
return true;
}
}
-23
View File
@@ -1,23 +0,0 @@
import { Global, Module } from '@nestjs/common';
import { createAuth, type Auth } from '@mosaic/auth';
import type { Db } from '@mosaic/db';
import { DB } from '../database/database.module.js';
import { AUTH } from './auth.tokens.js';
@Global()
@Module({
providers: [
{
provide: AUTH,
useFactory: (db: Db): Auth =>
createAuth({
db,
baseURL: process.env['BETTER_AUTH_URL'] ?? 'http://localhost:4000',
secret: process.env['BETTER_AUTH_SECRET'],
}),
inject: [DB],
},
],
exports: [AUTH],
})
export class AuthModule {}
-1
View File
@@ -1 +0,0 @@
export const AUTH = 'AUTH';
@@ -1,7 +0,0 @@
import { createParamDecorator, type ExecutionContext } from '@nestjs/common';
import type { FastifyRequest } from 'fastify';
export const CurrentUser = createParamDecorator((_data: unknown, ctx: ExecutionContext) => {
const request = ctx.switchToHttp().getRequest<FastifyRequest & { user?: unknown }>();
return request.user;
});
@@ -1,11 +0,0 @@
import { ForbiddenException } from '@nestjs/common';
export function assertOwner(
ownerId: string | null | undefined,
userId: string,
resourceName: string,
): void {
if (!ownerId || ownerId !== userId) {
throw new ForbiddenException(`${resourceName} does not belong to the current user`);
}
}
-18
View File
@@ -1,18 +0,0 @@
import { Global, Module } from '@nestjs/common';
import { createBrain, type Brain } from '@mosaic/brain';
import type { Db } from '@mosaic/db';
import { DB } from '../database/database.module.js';
import { BRAIN } from './brain.tokens.js';
@Global()
@Module({
providers: [
{
provide: BRAIN,
useFactory: (db: Db): Brain => createBrain(db),
inject: [DB],
},
],
exports: [BRAIN],
})
export class BrainModule {}
-1
View File
@@ -1 +0,0 @@
export const BRAIN = 'BRAIN';
@@ -1,80 +0,0 @@
import { readFileSync } from 'node:fs';
import { resolve } from 'node:path';
import { validateSync } from 'class-validator';
import { describe, expect, it, vi } from 'vitest';
import { SendMessageDto } from '../../conversations/conversations.dto.js';
import { ChatRequestDto } from '../chat.dto.js';
import { validateSocketSession } from '../chat.gateway-auth.js';
describe('Chat controller source hardening', () => {
it('applies AuthGuard and reads the current user', () => {
const source = readFileSync(resolve('src/chat/chat.controller.ts'), 'utf8');
expect(source).toContain('@UseGuards(AuthGuard)');
expect(source).toContain('@CurrentUser() user: { id: string }');
});
});
describe('WebSocket session authentication', () => {
it('returns null when the handshake does not resolve to a session', async () => {
const result = await validateSocketSession(
{},
{
api: {
getSession: vi.fn().mockResolvedValue(null),
},
},
);
expect(result).toBeNull();
});
it('returns the resolved session when Better Auth accepts the headers', async () => {
const session = { user: { id: 'user-1' }, session: { id: 'session-1' } };
const result = await validateSocketSession(
{ cookie: 'session=abc' },
{
api: {
getSession: vi.fn().mockResolvedValue(session),
},
},
);
expect(result).toEqual(session);
});
});
describe('Chat DTO validation', () => {
it('rejects unsupported message roles', () => {
const dto = Object.assign(new SendMessageDto(), {
content: 'hello',
role: 'moderator',
});
const errors = validateSync(dto);
expect(errors.length).toBeGreaterThan(0);
});
it('rejects oversized conversation message content above 10000 characters', () => {
const dto = Object.assign(new SendMessageDto(), {
content: 'x'.repeat(10_001),
role: 'user',
});
const errors = validateSync(dto);
expect(errors.length).toBeGreaterThan(0);
});
it('rejects oversized chat content above 10000 characters', () => {
const dto = Object.assign(new ChatRequestDto(), {
content: 'x'.repeat(10_001),
});
const errors = validateSync(dto);
expect(errors.length).toBeGreaterThan(0);
});
});
-90
View File
@@ -1,90 +0,0 @@
import {
Controller,
Post,
Body,
Logger,
HttpException,
HttpStatus,
Inject,
UseGuards,
} from '@nestjs/common';
import type { AgentSessionEvent } from '@mariozechner/pi-coding-agent';
import { Throttle } from '@nestjs/throttler';
import { AgentService } from '../agent/agent.service.js';
import { AuthGuard } from '../auth/auth.guard.js';
import { CurrentUser } from '../auth/current-user.decorator.js';
import { v4 as uuid } from 'uuid';
import { ChatRequestDto } from './chat.dto.js';
interface ChatResponse {
conversationId: string;
text: string;
}
@Controller('api/chat')
@UseGuards(AuthGuard)
export class ChatController {
private readonly logger = new Logger(ChatController.name);
constructor(@Inject(AgentService) private readonly agentService: AgentService) {}
@Post()
@Throttle({ default: { limit: 10, ttl: 60_000 } })
async chat(
@Body() body: ChatRequestDto,
@CurrentUser() user: { id: string },
): Promise<ChatResponse> {
const conversationId = body.conversationId ?? uuid();
try {
let agentSession = this.agentService.getSession(conversationId);
if (!agentSession) {
agentSession = await this.agentService.createSession(conversationId);
}
} catch (err) {
this.logger.error(
`Session creation failed for conversation=${conversationId}`,
err instanceof Error ? err.stack : String(err),
);
throw new HttpException('Agent session unavailable', HttpStatus.SERVICE_UNAVAILABLE);
}
this.logger.debug(`Handling chat request for user=${user.id}, conversation=${conversationId}`);
let responseText = '';
const done = new Promise<void>((resolve, reject) => {
const timer = setTimeout(() => {
cleanup();
this.logger.error(`Agent response timed out after 120s for conversation=${conversationId}`);
reject(new Error('Agent response timed out'));
}, 120_000);
const cleanup = this.agentService.onEvent(conversationId, (event: AgentSessionEvent) => {
if (event.type === 'message_update' && event.assistantMessageEvent.type === 'text_delta') {
responseText += event.assistantMessageEvent.delta;
}
if (event.type === 'agent_end') {
clearTimeout(timer);
cleanup();
resolve();
}
});
});
try {
await this.agentService.prompt(conversationId, body.content);
await done;
} catch (err) {
if (err instanceof HttpException) throw err;
const message = err instanceof Error ? err.message : String(err);
if (message.includes('timed out')) {
throw new HttpException('Agent response timed out', HttpStatus.GATEWAY_TIMEOUT);
}
this.logger.error(`Chat prompt failed for conversation=${conversationId}`, String(err));
throw new HttpException('Agent processing failed', HttpStatus.INTERNAL_SERVER_ERROR);
}
return { conversationId, text: responseText };
}
}
-35
View File
@@ -1,35 +0,0 @@
import { IsOptional, IsString, IsUUID, MaxLength } from 'class-validator';
export class ChatRequestDto {
@IsOptional()
@IsUUID()
conversationId?: string;
@IsString()
@MaxLength(10_000)
content!: string;
}
export class ChatSocketMessageDto {
@IsOptional()
@IsUUID()
conversationId?: string;
@IsString()
@MaxLength(10_000)
content!: string;
@IsOptional()
@IsString()
@MaxLength(255)
provider?: string;
@IsOptional()
@IsString()
@MaxLength(255)
modelId?: string;
@IsOptional()
@IsUUID()
agentId?: string;
}
@@ -1,30 +0,0 @@
import type { IncomingHttpHeaders } from 'node:http';
import { fromNodeHeaders } from 'better-auth/node';
export interface SocketSessionResult {
session: unknown;
user: { id: string };
}
export interface SessionAuth {
api: {
getSession(context: { headers: Headers }): Promise<SocketSessionResult | null>;
};
}
export async function validateSocketSession(
headers: IncomingHttpHeaders,
auth: SessionAuth,
): Promise<SocketSessionResult | null> {
const sessionHeaders = fromNodeHeaders(headers);
const result = await auth.api.getSession({ headers: sessionHeaders });
if (!result) {
return null;
}
return {
session: result.session,
user: { id: result.user.id },
};
}
-284
View File
@@ -1,284 +0,0 @@
import { Inject, Logger } from '@nestjs/common';
import {
WebSocketGateway,
WebSocketServer,
SubscribeMessage,
OnGatewayConnection,
OnGatewayDisconnect,
type OnGatewayInit,
ConnectedSocket,
MessageBody,
} from '@nestjs/websockets';
import { Server, Socket } from 'socket.io';
import type { AgentSessionEvent } from '@mariozechner/pi-coding-agent';
import type { Auth } from '@mosaic/auth';
import type { SetThinkingPayload, SlashCommandPayload, SystemReloadPayload } from '@mosaic/types';
import { AgentService } from '../agent/agent.service.js';
import { AUTH } from '../auth/auth.tokens.js';
import { CommandRegistryService } from '../commands/command-registry.service.js';
import { CommandExecutorService } from '../commands/command-executor.service.js';
import { v4 as uuid } from 'uuid';
import { ChatSocketMessageDto } from './chat.dto.js';
import { validateSocketSession } from './chat.gateway-auth.js';
@WebSocketGateway({
cors: {
origin: process.env['GATEWAY_CORS_ORIGIN'] ?? 'http://localhost:3000',
},
namespace: '/chat',
})
export class ChatGateway implements OnGatewayInit, OnGatewayConnection, OnGatewayDisconnect {
@WebSocketServer()
server!: Server;
private readonly logger = new Logger(ChatGateway.name);
private readonly clientSessions = new Map<
string,
{ conversationId: string; cleanup: () => void }
>();
constructor(
@Inject(AgentService) private readonly agentService: AgentService,
@Inject(AUTH) private readonly auth: Auth,
@Inject(CommandRegistryService) private readonly commandRegistry: CommandRegistryService,
@Inject(CommandExecutorService) private readonly commandExecutor: CommandExecutorService,
) {}
afterInit(): void {
this.logger.log('Chat WebSocket gateway initialized');
}
async handleConnection(client: Socket): Promise<void> {
const session = await validateSocketSession(client.handshake.headers, this.auth);
if (!session) {
this.logger.warn(`Rejected unauthenticated WebSocket client: ${client.id}`);
client.disconnect();
return;
}
client.data.user = session.user;
client.data.session = session.session;
this.logger.log(`Client connected: ${client.id}`);
// Broadcast command manifest to the newly connected client
client.emit('commands:manifest', { manifest: this.commandRegistry.getManifest() });
}
handleDisconnect(client: Socket): void {
this.logger.log(`Client disconnected: ${client.id}`);
const session = this.clientSessions.get(client.id);
if (session) {
session.cleanup();
this.agentService.removeChannel(session.conversationId, `websocket:${client.id}`);
this.clientSessions.delete(client.id);
}
}
@SubscribeMessage('message')
async handleMessage(
@ConnectedSocket() client: Socket,
@MessageBody() data: ChatSocketMessageDto,
): Promise<void> {
const conversationId = data.conversationId ?? uuid();
this.logger.log(`Message from ${client.id} in conversation ${conversationId}`);
// Ensure agent session exists for this conversation
try {
let agentSession = this.agentService.getSession(conversationId);
if (!agentSession) {
const userId = (client.data.user as { id: string } | undefined)?.id;
agentSession = await this.agentService.createSession(conversationId, {
provider: data.provider,
modelId: data.modelId,
agentConfigId: data.agentId,
userId,
});
}
} catch (err) {
this.logger.error(
`Session creation failed for client=${client.id}, conversation=${conversationId}`,
err instanceof Error ? err.stack : String(err),
);
client.emit('error', {
conversationId,
error: 'Failed to start agent session. Please try again.',
});
return;
}
// Always clean up previous listener to prevent leak
const existing = this.clientSessions.get(client.id);
if (existing) {
existing.cleanup();
}
// Subscribe to agent events and relay to client
const cleanup = this.agentService.onEvent(conversationId, (event: AgentSessionEvent) => {
this.relayEvent(client, conversationId, event);
});
this.clientSessions.set(client.id, { conversationId, cleanup });
// Track channel connection
this.agentService.addChannel(conversationId, `websocket:${client.id}`);
// Send session info so the client knows the model/provider
{
const agentSession = this.agentService.getSession(conversationId);
if (agentSession) {
const piSession = agentSession.piSession;
client.emit('session:info', {
conversationId,
provider: agentSession.provider,
modelId: agentSession.modelId,
thinkingLevel: piSession.thinkingLevel,
availableThinkingLevels: piSession.getAvailableThinkingLevels(),
});
}
}
// Send acknowledgment
client.emit('message:ack', { conversationId, messageId: uuid() });
// Dispatch to agent
try {
await this.agentService.prompt(conversationId, data.content);
} catch (err) {
this.logger.error(
`Agent prompt failed for client=${client.id}, conversation=${conversationId}`,
err instanceof Error ? err.stack : String(err),
);
client.emit('error', {
conversationId,
error: 'The agent failed to process your message. Please try again.',
});
}
}
@SubscribeMessage('set:thinking')
handleSetThinking(
@ConnectedSocket() client: Socket,
@MessageBody() data: SetThinkingPayload,
): void {
const session = this.agentService.getSession(data.conversationId);
if (!session) {
client.emit('error', {
conversationId: data.conversationId,
error: 'No active session for this conversation.',
});
return;
}
const validLevels = session.piSession.getAvailableThinkingLevels();
if (!validLevels.includes(data.level as never)) {
client.emit('error', {
conversationId: data.conversationId,
error: `Invalid thinking level "${data.level}". Available: ${validLevels.join(', ')}`,
});
return;
}
session.piSession.setThinkingLevel(data.level as never);
this.logger.log(
`Thinking level set to "${data.level}" for conversation ${data.conversationId}`,
);
client.emit('session:info', {
conversationId: data.conversationId,
provider: session.provider,
modelId: session.modelId,
thinkingLevel: session.piSession.thinkingLevel,
availableThinkingLevels: session.piSession.getAvailableThinkingLevels(),
});
}
@SubscribeMessage('command:execute')
async handleCommandExecute(
@ConnectedSocket() client: Socket,
@MessageBody() payload: SlashCommandPayload,
): Promise<void> {
const userId = (client.data.user as { id: string } | undefined)?.id ?? 'unknown';
const result = await this.commandExecutor.execute(payload, userId);
client.emit('command:result', result);
}
broadcastReload(payload: SystemReloadPayload): void {
this.server.emit('system:reload', payload);
this.logger.log('Broadcasted system:reload to all connected clients');
}
private relayEvent(client: Socket, conversationId: string, event: AgentSessionEvent): void {
if (!client.connected) {
this.logger.warn(
`Dropping event ${event.type} for disconnected client=${client.id}, conversation=${conversationId}`,
);
return;
}
switch (event.type) {
case 'agent_start':
client.emit('agent:start', { conversationId });
break;
case 'agent_end': {
// Gather usage stats from the Pi session
const agentSession = this.agentService.getSession(conversationId);
const piSession = agentSession?.piSession;
const stats = piSession?.getSessionStats();
const contextUsage = piSession?.getContextUsage();
client.emit('agent:end', {
conversationId,
usage: stats
? {
provider: agentSession?.provider ?? 'unknown',
modelId: agentSession?.modelId ?? 'unknown',
thinkingLevel: piSession?.thinkingLevel ?? 'off',
tokens: stats.tokens,
cost: stats.cost,
context: {
percent: contextUsage?.percent ?? null,
window: contextUsage?.contextWindow ?? 0,
},
}
: undefined,
});
break;
}
case 'message_update': {
const assistantEvent = event.assistantMessageEvent;
if (assistantEvent.type === 'text_delta') {
client.emit('agent:text', {
conversationId,
text: assistantEvent.delta,
});
} else if (assistantEvent.type === 'thinking_delta') {
client.emit('agent:thinking', {
conversationId,
text: assistantEvent.delta,
});
}
break;
}
case 'tool_execution_start':
client.emit('agent:tool:start', {
conversationId,
toolCallId: event.toolCallId,
toolName: event.toolName,
});
break;
case 'tool_execution_end':
client.emit('agent:tool:end', {
conversationId,
toolCallId: event.toolCallId,
toolName: event.toolName,
isError: event.isError,
});
break;
}
}
}
-12
View File
@@ -1,12 +0,0 @@
import { forwardRef, Module } from '@nestjs/common';
import { CommandsModule } from '../commands/commands.module.js';
import { ChatGateway } from './chat.gateway.js';
import { ChatController } from './chat.controller.js';
@Module({
imports: [forwardRef(() => CommandsModule)],
controllers: [ChatController],
providers: [ChatGateway],
exports: [ChatGateway],
})
export class ChatModule {}
@@ -1,213 +0,0 @@
import { describe, it, expect, vi, beforeEach } from 'vitest';
import { CommandExecutorService } from './command-executor.service.js';
import type { SlashCommandPayload } from '@mosaic/types';
// Minimal mock implementations
const mockRegistry = {
getManifest: vi.fn(() => ({
version: 1,
commands: [
{ name: 'provider', aliases: [], scope: 'agent', execution: 'hybrid', available: true },
{ name: 'mission', aliases: [], scope: 'agent', execution: 'socket', available: true },
{ name: 'agent', aliases: ['a'], scope: 'agent', execution: 'socket', available: true },
{ name: 'prdy', aliases: [], scope: 'agent', execution: 'socket', available: true },
{ name: 'tools', aliases: [], scope: 'agent', execution: 'socket', available: true },
],
skills: [],
})),
};
const mockAgentService = {
getSession: vi.fn(() => undefined),
};
const mockSystemOverride = {
set: vi.fn(),
get: vi.fn(),
clear: vi.fn(),
renew: vi.fn(),
};
const mockSessionGC = {
sweepOrphans: vi.fn(() => ({ orphanedSessions: 0, totalCleaned: [], duration: 0 })),
};
const mockRedis = {
set: vi.fn().mockResolvedValue('OK'),
get: vi.fn(),
del: vi.fn(),
};
function buildService(): CommandExecutorService {
return new CommandExecutorService(
mockRegistry as never,
mockAgentService as never,
mockSystemOverride as never,
mockSessionGC as never,
mockRedis as never,
null,
null,
);
}
describe('CommandExecutorService — P8-012 commands', () => {
let service: CommandExecutorService;
const userId = 'user-123';
const conversationId = 'conv-456';
beforeEach(() => {
vi.clearAllMocks();
service = buildService();
});
// /provider login — missing provider name
it('/provider login with no provider name returns usage error', async () => {
const payload: SlashCommandPayload = { command: 'provider', args: 'login', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(false);
expect(result.message).toContain('Usage: /provider login');
expect(result.command).toBe('provider');
});
// /provider login anthropic — success with URL containing poll token
it('/provider login <name> returns success with URL and poll token', async () => {
const payload: SlashCommandPayload = {
command: 'provider',
args: 'login anthropic',
conversationId,
};
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('provider');
expect(result.message).toContain('anthropic');
expect(result.message).toContain('http');
// data should contain loginUrl and pollToken
expect(result.data).toBeDefined();
const data = result.data as Record<string, unknown>;
expect(typeof data['loginUrl']).toBe('string');
expect(typeof data['pollToken']).toBe('string');
expect(data['loginUrl'] as string).toContain('anthropic');
expect(data['loginUrl'] as string).toContain(data['pollToken'] as string);
// Verify Valkey was called
expect(mockRedis.set).toHaveBeenCalledOnce();
const [key, value, , ttl] = mockRedis.set.mock.calls[0] as [string, string, string, number];
expect(key).toContain('mosaic:auth:poll:');
const stored = JSON.parse(value) as { status: string; provider: string; userId: string };
expect(stored.status).toBe('pending');
expect(stored.provider).toBe('anthropic');
expect(stored.userId).toBe(userId);
expect(ttl).toBe(300);
});
// /provider with no args — returns usage
it('/provider with no args returns usage message', async () => {
const payload: SlashCommandPayload = { command: 'provider', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('Usage: /provider');
});
// /provider list
it('/provider list returns success', async () => {
const payload: SlashCommandPayload = { command: 'provider', args: 'list', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('provider');
});
// /provider logout with no name — usage error
it('/provider logout with no name returns error', async () => {
const payload: SlashCommandPayload = { command: 'provider', args: 'logout', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(false);
expect(result.message).toContain('Usage: /provider logout');
});
// /provider unknown subcommand
it('/provider unknown subcommand returns error', async () => {
const payload: SlashCommandPayload = {
command: 'provider',
args: 'unknown',
conversationId,
};
const result = await service.execute(payload, userId);
expect(result.success).toBe(false);
expect(result.message).toContain('Unknown subcommand');
});
// /mission status
it('/mission status returns stub message', async () => {
const payload: SlashCommandPayload = { command: 'mission', args: 'status', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('mission');
expect(result.message).toContain('Mission status');
});
// /mission with no args
it('/mission with no args returns status stub', async () => {
const payload: SlashCommandPayload = { command: 'mission', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('Mission status');
});
// /mission set <id>
it('/mission set <id> returns confirmation', async () => {
const payload: SlashCommandPayload = {
command: 'mission',
args: 'set my-mission-123',
conversationId,
};
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('my-mission-123');
});
// /agent list
it('/agent list returns stub message', async () => {
const payload: SlashCommandPayload = { command: 'agent', args: 'list', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('agent');
expect(result.message).toContain('agent');
});
// /agent with no args
it('/agent with no args returns usage', async () => {
const payload: SlashCommandPayload = { command: 'agent', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('Usage: /agent');
});
// /agent <id> — switch
it('/agent <id> returns switch confirmation', async () => {
const payload: SlashCommandPayload = {
command: 'agent',
args: 'my-agent-id',
conversationId,
};
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('my-agent-id');
});
// /prdy
it('/prdy returns PRD wizard message', async () => {
const payload: SlashCommandPayload = { command: 'prdy', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('prdy');
expect(result.message).toContain('mosaic prdy');
});
// /tools
it('/tools returns tools stub message', async () => {
const payload: SlashCommandPayload = { command: 'tools', conversationId };
const result = await service.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('tools');
expect(result.message).toContain('tools');
});
});
@@ -1,373 +0,0 @@
import { forwardRef, Inject, Injectable, Logger, Optional } from '@nestjs/common';
import type { QueueHandle } from '@mosaic/queue';
import type { SlashCommandPayload, SlashCommandResultPayload } from '@mosaic/types';
import { AgentService } from '../agent/agent.service.js';
import { ChatGateway } from '../chat/chat.gateway.js';
import { SessionGCService } from '../gc/session-gc.service.js';
import { SystemOverrideService } from '../preferences/system-override.service.js';
import { ReloadService } from '../reload/reload.service.js';
import { COMMANDS_REDIS } from './commands.tokens.js';
import { CommandRegistryService } from './command-registry.service.js';
@Injectable()
export class CommandExecutorService {
private readonly logger = new Logger(CommandExecutorService.name);
constructor(
@Inject(CommandRegistryService) private readonly registry: CommandRegistryService,
@Inject(AgentService) private readonly agentService: AgentService,
@Inject(SystemOverrideService) private readonly systemOverride: SystemOverrideService,
@Inject(SessionGCService) private readonly sessionGC: SessionGCService,
@Inject(COMMANDS_REDIS) private readonly redis: QueueHandle['redis'],
@Optional()
@Inject(forwardRef(() => ReloadService))
private readonly reloadService: ReloadService | null,
@Optional()
@Inject(forwardRef(() => ChatGateway))
private readonly chatGateway: ChatGateway | null,
) {}
async execute(payload: SlashCommandPayload, userId: string): Promise<SlashCommandResultPayload> {
const { command, args, conversationId } = payload;
const def = this.registry.getManifest().commands.find((c) => c.name === command);
if (!def) {
return {
command,
conversationId,
success: false,
message: `Unknown command: /${command}`,
};
}
try {
switch (command) {
case 'model':
return await this.handleModel(args ?? null, conversationId);
case 'thinking':
return await this.handleThinking(args ?? null, conversationId);
case 'system':
return await this.handleSystem(args ?? null, conversationId);
case 'new':
return {
command,
conversationId,
success: true,
message: 'Start a new conversation by selecting New Conversation.',
};
case 'clear':
return {
command,
conversationId,
success: true,
message: 'Conversation display cleared.',
};
case 'compact':
return {
command,
conversationId,
success: true,
message: 'Context compaction requested.',
};
case 'retry':
return {
command,
conversationId,
success: true,
message: 'Retry last message requested.',
};
case 'gc': {
// User-scoped sweep for non-admin; system-wide for admin
const result = await this.sessionGC.sweepOrphans(userId);
return {
command: 'gc',
success: true,
message: `GC sweep complete: ${result.orphanedSessions} orphaned sessions cleaned in ${result.duration}ms.`,
conversationId,
};
}
case 'agent':
return await this.handleAgent(args ?? null, conversationId);
case 'provider':
return await this.handleProvider(args ?? null, userId, conversationId);
case 'mission':
return await this.handleMission(args ?? null, conversationId, userId);
case 'prdy':
return {
command: 'prdy',
success: true,
message:
'PRD wizard: run `mosaic prdy` in your project workspace to create or update a PRD.',
conversationId,
};
case 'tools':
return await this.handleTools(conversationId, userId);
case 'reload': {
if (!this.reloadService) {
return {
command: 'reload',
conversationId,
success: false,
message: 'ReloadService is not available.',
};
}
const reloadResult = await this.reloadService.reload('command');
this.chatGateway?.broadcastReload(reloadResult);
return {
command: 'reload',
success: true,
message: reloadResult.message,
conversationId,
};
}
default:
return {
command,
conversationId,
success: false,
message: `Command /${command} is not yet implemented.`,
};
}
} catch (err) {
this.logger.error(`Command /${command} failed: ${err}`);
return { command, conversationId, success: false, message: String(err) };
}
}
private async handleModel(
args: string | null,
conversationId: string,
): Promise<SlashCommandResultPayload> {
if (!args) {
return {
command: 'model',
conversationId,
success: true,
message: 'Usage: /model <model-name>',
};
}
// Update agent session model if session is active
// For now, acknowledge the request — full wiring done in P8-012
const session = this.agentService.getSession(conversationId);
if (!session) {
return {
command: 'model',
conversationId,
success: true,
message: `Model switch to "${args}" requested. No active session for this conversation.`,
};
}
return {
command: 'model',
conversationId,
success: true,
message: `Model switch to "${args}" requested.`,
};
}
private async handleThinking(
args: string | null,
conversationId: string,
): Promise<SlashCommandResultPayload> {
const level = args?.toLowerCase();
if (!level || !['none', 'low', 'medium', 'high', 'auto'].includes(level)) {
return {
command: 'thinking',
conversationId,
success: true,
message: 'Usage: /thinking <none|low|medium|high|auto>',
};
}
return {
command: 'thinking',
conversationId,
success: true,
message: `Thinking level set to "${level}".`,
};
}
private async handleSystem(
args: string | null,
conversationId: string,
): Promise<SlashCommandResultPayload> {
if (!args || args.trim().length === 0) {
// Clear the override when called with no args
await this.systemOverride.clear(conversationId);
return {
command: 'system',
conversationId,
success: true,
message: 'Session system prompt override cleared.',
};
}
await this.systemOverride.set(conversationId, args.trim());
return {
command: 'system',
conversationId,
success: true,
message: `Session system prompt override set (expires in 5 minutes of inactivity).`,
};
}
private async handleAgent(
args: string | null,
conversationId: string,
): Promise<SlashCommandResultPayload> {
if (!args) {
return {
command: 'agent',
success: true,
message: 'Usage: /agent <agent-id> to switch, or /agent list to see available agents.',
conversationId,
};
}
if (args === 'list') {
return {
command: 'agent',
success: true,
message: 'Agent listing: use the web dashboard for full agent management.',
conversationId,
};
}
// Switch agent — stub for now (full implementation in P8-015)
return {
command: 'agent',
success: true,
message: `Agent switch to "${args}" requested. Restart conversation to apply.`,
conversationId,
};
}
private async handleProvider(
args: string | null,
userId: string,
conversationId: string,
): Promise<SlashCommandResultPayload> {
if (!args) {
return {
command: 'provider',
success: true,
message: 'Usage: /provider list | /provider login <name> | /provider logout <name>',
conversationId,
};
}
const spaceIdx = args.indexOf(' ');
const subcommand = spaceIdx >= 0 ? args.slice(0, spaceIdx) : args;
const providerName = spaceIdx >= 0 ? args.slice(spaceIdx + 1).trim() : '';
switch (subcommand) {
case 'list':
return {
command: 'provider',
success: true,
message: 'Use the web dashboard to manage providers.',
conversationId,
};
case 'login': {
if (!providerName) {
return {
command: 'provider',
success: false,
message: 'Usage: /provider login <provider-name>',
conversationId,
};
}
const pollToken = crypto.randomUUID();
const key = `mosaic:auth:poll:${pollToken}`;
// Store pending state in Valkey (TTL 5 minutes)
await this.redis.set(
key,
JSON.stringify({ status: 'pending', provider: providerName, userId }),
'EX',
300,
);
// In production this would construct an OAuth URL
const loginUrl = `${process.env['MOSAIC_BASE_URL'] ?? 'http://localhost:3000'}/auth/provider/${providerName}?token=${pollToken}`;
return {
command: 'provider',
success: true,
message: `Open this URL to authenticate with ${providerName}:\n${loginUrl}`,
conversationId,
data: { loginUrl, pollToken, provider: providerName },
};
}
case 'logout': {
if (!providerName) {
return {
command: 'provider',
success: false,
message: 'Usage: /provider logout <provider-name>',
conversationId,
};
}
return {
command: 'provider',
success: true,
message: `Logout from ${providerName}: use the web dashboard to revoke provider tokens.`,
conversationId,
};
}
default:
return {
command: 'provider',
success: false,
message: `Unknown subcommand: ${subcommand}. Use list, login, or logout.`,
conversationId,
};
}
}
private async handleMission(
args: string | null,
conversationId: string,
_userId: string,
): Promise<SlashCommandResultPayload> {
if (!args || args === 'status') {
// TODO: fetch active mission from DB when MissionsService is available
return {
command: 'mission',
success: true,
message: 'Mission status: use the web dashboard for full mission management.',
conversationId,
};
}
if (args.startsWith('set ')) {
const missionId = args.slice(4).trim();
return {
command: 'mission',
success: true,
message: `Mission set to ${missionId}. Session context updated.`,
conversationId,
};
}
return {
command: 'mission',
success: true,
message: 'Usage: /mission [status|set <id>|list|tasks]',
conversationId,
};
}
private async handleTools(
conversationId: string,
_userId: string,
): Promise<SlashCommandResultPayload> {
// TODO: fetch tool list from active agent session
return {
command: 'tools',
success: true,
message:
'Available tools depend on the active agent configuration. Use the web dashboard to configure tool access.',
conversationId,
};
}
}
@@ -1,53 +0,0 @@
import { describe, it, expect, beforeEach } from 'vitest';
import { CommandRegistryService } from './command-registry.service.js';
import type { CommandDef } from '@mosaic/types';
const mockCmd: CommandDef = {
name: 'test',
description: 'Test command',
aliases: ['t'],
scope: 'core',
execution: 'local',
available: true,
};
describe('CommandRegistryService', () => {
let service: CommandRegistryService;
beforeEach(() => {
service = new CommandRegistryService();
});
it('starts with empty manifest', () => {
expect(service.getManifest().commands).toHaveLength(0);
});
it('registers a command', () => {
service.registerCommand(mockCmd);
expect(service.getManifest().commands).toHaveLength(1);
});
it('updates existing command by name', () => {
service.registerCommand(mockCmd);
service.registerCommand({ ...mockCmd, description: 'Updated' });
expect(service.getManifest().commands).toHaveLength(1);
expect(service.getManifest().commands[0]?.description).toBe('Updated');
});
it('onModuleInit registers core commands', () => {
service.onModuleInit();
const manifest = service.getManifest();
expect(manifest.commands.length).toBeGreaterThan(5);
expect(manifest.commands.some((c) => c.name === 'model')).toBe(true);
expect(manifest.commands.some((c) => c.name === 'help')).toBe(true);
});
it('manifest includes skills array', () => {
const manifest = service.getManifest();
expect(Array.isArray(manifest.skills)).toBe(true);
});
it('manifest version is 1', () => {
expect(service.getManifest().version).toBe(1);
});
});
@@ -1,273 +0,0 @@
import { Injectable, type OnModuleInit } from '@nestjs/common';
import type { CommandDef, CommandManifest } from '@mosaic/types';
@Injectable()
export class CommandRegistryService implements OnModuleInit {
private readonly commands: CommandDef[] = [];
registerCommand(def: CommandDef): void {
const existing = this.commands.findIndex((c) => c.name === def.name);
if (existing >= 0) {
this.commands[existing] = def;
} else {
this.commands.push(def);
}
}
registerCommands(defs: CommandDef[]): void {
for (const def of defs) {
this.registerCommand(def);
}
}
getManifest(): CommandManifest {
return {
version: 1,
commands: [...this.commands],
skills: [],
};
}
onModuleInit(): void {
this.registerCommands([
{
name: 'model',
description: 'Switch the active model',
aliases: ['m'],
args: [
{
name: 'model-name',
type: 'string',
optional: false,
description: 'Model name to switch to',
},
],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'thinking',
description: 'Set thinking level (none/low/medium/high/auto)',
aliases: ['t'],
args: [
{
name: 'level',
type: 'enum',
optional: false,
values: ['none', 'low', 'medium', 'high', 'auto'],
description: 'Thinking level',
},
],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'new',
description: 'Start a new conversation',
aliases: ['n'],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'clear',
description: 'Clear conversation context and GC session artifacts',
aliases: [],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'compact',
description: 'Request context compaction',
aliases: [],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'retry',
description: 'Retry the last message',
aliases: [],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'rename',
description: 'Rename current conversation',
aliases: [],
args: [
{ name: 'name', type: 'string', optional: false, description: 'New conversation name' },
],
scope: 'core',
execution: 'rest',
available: true,
},
{
name: 'history',
description: 'Show conversation history',
aliases: [],
args: [
{
name: 'limit',
type: 'string',
optional: true,
description: 'Number of messages to show',
},
],
scope: 'core',
execution: 'rest',
available: true,
},
{
name: 'export',
description: 'Export conversation to markdown or JSON',
aliases: [],
args: [
{
name: 'format',
type: 'enum',
optional: true,
values: ['md', 'json'],
description: 'Export format',
},
],
scope: 'core',
execution: 'rest',
available: true,
},
{
name: 'preferences',
description: 'View or set user preferences',
aliases: ['pref'],
args: [
{
name: 'action',
type: 'enum',
optional: true,
values: ['show', 'set', 'reset'],
description: 'Action to perform',
},
],
scope: 'core',
execution: 'rest',
available: true,
},
{
name: 'system',
description: 'Set session-scoped system prompt override',
aliases: [],
args: [
{
name: 'override',
type: 'string',
optional: false,
description: 'System prompt text to inject for this session',
},
],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'status',
description: 'Show session and connection status',
aliases: ['s'],
scope: 'core',
execution: 'hybrid',
available: true,
},
{
name: 'help',
description: 'Show available commands',
aliases: ['h'],
scope: 'core',
execution: 'local',
available: true,
},
{
name: 'gc',
description: 'Trigger garbage collection sweep (user-scoped)',
aliases: [],
scope: 'core',
execution: 'socket',
available: true,
},
{
name: 'agent',
description: 'Switch or list available agents',
aliases: ['a'],
args: [
{
name: 'args',
type: 'string',
optional: true,
description: 'list or <agent-id>',
},
],
scope: 'agent',
execution: 'socket',
available: true,
},
{
name: 'provider',
description: 'Manage LLM providers (list/login/logout)',
aliases: [],
args: [
{
name: 'args',
type: 'string',
optional: true,
description: 'list | login <name> | logout <name>',
},
],
scope: 'agent',
execution: 'hybrid',
available: true,
},
{
name: 'mission',
description: 'View or set active mission',
aliases: [],
args: [
{
name: 'args',
type: 'string',
optional: true,
description: 'status | set <id> | list | tasks',
},
],
scope: 'agent',
execution: 'socket',
available: true,
},
{
name: 'prdy',
description: 'Launch PRD wizard',
aliases: [],
scope: 'agent',
execution: 'socket',
available: true,
},
{
name: 'tools',
description: 'List available agent tools',
aliases: [],
scope: 'agent',
execution: 'socket',
available: true,
},
{
name: 'reload',
description: 'Soft-reload gateway plugins and command manifest (admin)',
aliases: [],
scope: 'admin',
execution: 'socket',
available: true,
},
]);
}
}
@@ -1,253 +0,0 @@
/**
* Integration tests for the gateway command system (P8-019)
*
* Covers:
* - CommandRegistryService.getManifest() returns 12+ core commands
* - All core commands have correct execution types
* - Alias resolution works for all defined aliases
* - CommandExecutorService routes known/unknown commands correctly
* - /gc handler calls SessionGCService.sweepOrphans
* - /system handler calls SystemOverrideService.set
* - Unknown command returns descriptive error
*/
import { describe, it, expect, vi, beforeEach } from 'vitest';
import { CommandRegistryService } from './command-registry.service.js';
import { CommandExecutorService } from './command-executor.service.js';
import type { SlashCommandPayload } from '@mosaic/types';
// ─── Mocks ───────────────────────────────────────────────────────────────────
const mockAgentService = {
getSession: vi.fn(() => undefined),
};
const mockSystemOverride = {
set: vi.fn().mockResolvedValue(undefined),
get: vi.fn().mockResolvedValue(null),
clear: vi.fn().mockResolvedValue(undefined),
renew: vi.fn().mockResolvedValue(undefined),
};
const mockSessionGC = {
sweepOrphans: vi.fn().mockResolvedValue({ orphanedSessions: 3, totalCleaned: [], duration: 12 }),
};
const mockRedis = {
set: vi.fn().mockResolvedValue('OK'),
get: vi.fn().mockResolvedValue(null),
del: vi.fn().mockResolvedValue(0),
keys: vi.fn().mockResolvedValue([]),
};
// ─── Helpers ─────────────────────────────────────────────────────────────────
function buildRegistry(): CommandRegistryService {
const svc = new CommandRegistryService();
svc.onModuleInit(); // seed core commands
return svc;
}
function buildExecutor(registry: CommandRegistryService): CommandExecutorService {
return new CommandExecutorService(
registry as never,
mockAgentService as never,
mockSystemOverride as never,
mockSessionGC as never,
mockRedis as never,
null, // reloadService (optional)
null, // chatGateway (optional)
);
}
// ─── Registry Tests ───────────────────────────────────────────────────────────
describe('CommandRegistryService — integration', () => {
let registry: CommandRegistryService;
beforeEach(() => {
registry = buildRegistry();
});
it('getManifest() returns 12 or more core commands after onModuleInit', () => {
const manifest = registry.getManifest();
expect(manifest.commands.length).toBeGreaterThanOrEqual(12);
});
it('manifest version is 1', () => {
expect(registry.getManifest().version).toBe(1);
});
it('manifest.skills is an array', () => {
expect(Array.isArray(registry.getManifest().skills)).toBe(true);
});
it('all commands have required fields: name, description, execution, scope, available', () => {
for (const cmd of registry.getManifest().commands) {
expect(typeof cmd.name).toBe('string');
expect(typeof cmd.description).toBe('string');
expect(['local', 'socket', 'rest', 'hybrid']).toContain(cmd.execution);
expect(['core', 'agent', 'admin']).toContain(cmd.scope);
expect(typeof cmd.available).toBe('boolean');
}
});
// Execution type verification for core commands
const expectedExecutionTypes: Record<string, string> = {
model: 'socket',
thinking: 'socket',
new: 'socket',
clear: 'socket',
compact: 'socket',
retry: 'socket',
rename: 'rest',
history: 'rest',
export: 'rest',
preferences: 'rest',
system: 'socket',
help: 'local',
gc: 'socket',
agent: 'socket',
provider: 'hybrid',
mission: 'socket',
prdy: 'socket',
tools: 'socket',
reload: 'socket',
};
for (const [name, expectedExecution] of Object.entries(expectedExecutionTypes)) {
it(`command "${name}" has execution type "${expectedExecution}"`, () => {
const cmd = registry.getManifest().commands.find((c) => c.name === name);
expect(cmd, `command "${name}" not found`).toBeDefined();
expect(cmd!.execution).toBe(expectedExecution);
});
}
// Alias resolution checks
const expectedAliases: Array<[string, string]> = [
['m', 'model'],
['t', 'thinking'],
['n', 'new'],
['a', 'agent'],
['s', 'status'],
['h', 'help'],
['pref', 'preferences'],
];
for (const [alias, commandName] of expectedAliases) {
it(`alias "/${alias}" resolves to command "${commandName}" via aliases array`, () => {
const cmd = registry
.getManifest()
.commands.find((c) => c.name === commandName || c.aliases?.includes(alias));
expect(cmd, `command with alias "${alias}" not found`).toBeDefined();
});
}
});
// ─── Executor Tests ───────────────────────────────────────────────────────────
describe('CommandExecutorService — integration', () => {
let registry: CommandRegistryService;
let executor: CommandExecutorService;
const userId = 'user-integ-001';
const conversationId = 'conv-integ-001';
beforeEach(() => {
vi.clearAllMocks();
registry = buildRegistry();
executor = buildExecutor(registry);
});
// Unknown command returns error
it('unknown command returns success:false with descriptive message', async () => {
const payload: SlashCommandPayload = { command: 'nonexistent', conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(false);
expect(result.message).toContain('nonexistent');
expect(result.command).toBe('nonexistent');
});
// /gc handler calls SessionGCService.sweepOrphans
it('/gc calls SessionGCService.sweepOrphans with userId', async () => {
const payload: SlashCommandPayload = { command: 'gc', conversationId };
const result = await executor.execute(payload, userId);
expect(mockSessionGC.sweepOrphans).toHaveBeenCalledWith(userId);
expect(result.success).toBe(true);
expect(result.message).toContain('GC sweep complete');
expect(result.message).toContain('3 orphaned sessions');
});
// /system with args calls SystemOverrideService.set
it('/system with text calls SystemOverrideService.set', async () => {
const override = 'You are a helpful assistant.';
const payload: SlashCommandPayload = { command: 'system', args: override, conversationId };
const result = await executor.execute(payload, userId);
expect(mockSystemOverride.set).toHaveBeenCalledWith(conversationId, override);
expect(result.success).toBe(true);
expect(result.message).toContain('override set');
});
// /system with no args clears the override
it('/system with no args calls SystemOverrideService.clear', async () => {
const payload: SlashCommandPayload = { command: 'system', conversationId };
const result = await executor.execute(payload, userId);
expect(mockSystemOverride.clear).toHaveBeenCalledWith(conversationId);
expect(result.success).toBe(true);
expect(result.message).toContain('cleared');
});
// /model with model name returns success
it('/model with a model name returns success', async () => {
const payload: SlashCommandPayload = {
command: 'model',
args: 'claude-3-opus',
conversationId,
};
const result = await executor.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('model');
expect(result.message).toContain('claude-3-opus');
});
// /thinking with valid level returns success
it('/thinking with valid level returns success', async () => {
const payload: SlashCommandPayload = { command: 'thinking', args: 'high', conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('high');
});
// /thinking with invalid level returns usage message
it('/thinking with invalid level returns usage message', async () => {
const payload: SlashCommandPayload = { command: 'thinking', args: 'invalid', conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.message).toContain('Usage:');
});
// /new command returns success
it('/new returns success', async () => {
const payload: SlashCommandPayload = { command: 'new', conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe('new');
});
// /reload without reloadService returns failure
it('/reload without ReloadService returns failure', async () => {
const payload: SlashCommandPayload = { command: 'reload', conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(false);
expect(result.message).toContain('ReloadService');
});
// Commands not yet fully implemented return a fallback response
const stubCommands = ['clear', 'compact', 'retry'];
for (const cmd of stubCommands) {
it(`/${cmd} returns success (stub)`, async () => {
const payload: SlashCommandPayload = { command: cmd, conversationId };
const result = await executor.execute(payload, userId);
expect(result.success).toBe(true);
expect(result.command).toBe(cmd);
});
}
});
@@ -1,37 +0,0 @@
import { forwardRef, Inject, Module, type OnApplicationShutdown } from '@nestjs/common';
import { createQueue, type QueueHandle } from '@mosaic/queue';
import { ChatModule } from '../chat/chat.module.js';
import { GCModule } from '../gc/gc.module.js';
import { ReloadModule } from '../reload/reload.module.js';
import { CommandExecutorService } from './command-executor.service.js';
import { CommandRegistryService } from './command-registry.service.js';
import { COMMANDS_REDIS } from './commands.tokens.js';
const COMMANDS_QUEUE_HANDLE = 'COMMANDS_QUEUE_HANDLE';
@Module({
imports: [GCModule, forwardRef(() => ReloadModule), forwardRef(() => ChatModule)],
providers: [
{
provide: COMMANDS_QUEUE_HANDLE,
useFactory: (): QueueHandle => {
return createQueue();
},
},
{
provide: COMMANDS_REDIS,
useFactory: (handle: QueueHandle) => handle.redis,
inject: [COMMANDS_QUEUE_HANDLE],
},
CommandRegistryService,
CommandExecutorService,
],
exports: [CommandRegistryService, CommandExecutorService],
})
export class CommandsModule implements OnApplicationShutdown {
constructor(@Inject(COMMANDS_QUEUE_HANDLE) private readonly handle: QueueHandle) {}
async onApplicationShutdown(): Promise<void> {
await this.handle.close().catch(() => {});
}
}
@@ -1 +0,0 @@
export const COMMANDS_REDIS = 'COMMANDS_REDIS';
@@ -1,97 +0,0 @@
import {
Body,
Controller,
Delete,
Get,
HttpCode,
HttpStatus,
Inject,
NotFoundException,
Param,
Patch,
Post,
UseGuards,
} from '@nestjs/common';
import type { Brain } from '@mosaic/brain';
import { BRAIN } from '../brain/brain.tokens.js';
import { AuthGuard } from '../auth/auth.guard.js';
import { CurrentUser } from '../auth/current-user.decorator.js';
import { assertOwner } from '../auth/resource-ownership.js';
import {
CreateConversationDto,
UpdateConversationDto,
SendMessageDto,
} from './conversations.dto.js';
@Controller('api/conversations')
@UseGuards(AuthGuard)
export class ConversationsController {
constructor(@Inject(BRAIN) private readonly brain: Brain) {}
@Get()
async list(@CurrentUser() user: { id: string }) {
return this.brain.conversations.findAll(user.id);
}
@Get(':id')
async findOne(@Param('id') id: string, @CurrentUser() user: { id: string }) {
return this.getOwnedConversation(id, user.id);
}
@Post()
async create(@CurrentUser() user: { id: string }, @Body() dto: CreateConversationDto) {
return this.brain.conversations.create({
userId: user.id,
title: dto.title,
projectId: dto.projectId,
});
}
@Patch(':id')
async update(
@Param('id') id: string,
@Body() dto: UpdateConversationDto,
@CurrentUser() user: { id: string },
) {
await this.getOwnedConversation(id, user.id);
const conversation = await this.brain.conversations.update(id, dto);
if (!conversation) throw new NotFoundException('Conversation not found');
return conversation;
}
@Delete(':id')
@HttpCode(HttpStatus.NO_CONTENT)
async remove(@Param('id') id: string, @CurrentUser() user: { id: string }) {
await this.getOwnedConversation(id, user.id);
const deleted = await this.brain.conversations.remove(id);
if (!deleted) throw new NotFoundException('Conversation not found');
}
@Get(':id/messages')
async listMessages(@Param('id') id: string, @CurrentUser() user: { id: string }) {
await this.getOwnedConversation(id, user.id);
return this.brain.conversations.findMessages(id);
}
@Post(':id/messages')
async addMessage(
@Param('id') id: string,
@Body() dto: SendMessageDto,
@CurrentUser() user: { id: string },
) {
await this.getOwnedConversation(id, user.id);
return this.brain.conversations.addMessage({
conversationId: id,
role: dto.role,
content: dto.content,
metadata: dto.metadata,
});
}
private async getOwnedConversation(id: string, userId: string) {
const conversation = await this.brain.conversations.findById(id);
if (!conversation) throw new NotFoundException('Conversation not found');
assertOwner(conversation.userId, userId, 'Conversation');
return conversation;
}
}
@@ -1,48 +0,0 @@
import {
IsBoolean,
IsIn,
IsObject,
IsOptional,
IsString,
IsUUID,
MaxLength,
} from 'class-validator';
export class CreateConversationDto {
@IsOptional()
@IsString()
@MaxLength(255)
title?: string;
@IsOptional()
@IsUUID()
projectId?: string;
}
export class UpdateConversationDto {
@IsOptional()
@IsString()
@MaxLength(255)
title?: string;
@IsOptional()
@IsUUID()
projectId?: string | null;
@IsOptional()
@IsBoolean()
archived?: boolean;
}
export class SendMessageDto {
@IsIn(['user', 'assistant', 'system'])
role!: 'user' | 'assistant' | 'system';
@IsString()
@MaxLength(10_000)
content!: string;
@IsOptional()
@IsObject()
metadata?: Record<string, unknown>;
}
@@ -1,7 +0,0 @@
import { Module } from '@nestjs/common';
import { ConversationsController } from './conversations.controller.js';
@Module({
controllers: [ConversationsController],
})
export class ConversationsModule {}
@@ -1,77 +0,0 @@
import {
BadRequestException,
Controller,
Get,
Inject,
NotFoundException,
Param,
Query,
UseGuards,
} from '@nestjs/common';
import fs from 'node:fs';
import path from 'node:path';
import { AuthGuard } from '../auth/auth.guard.js';
import { CoordService } from './coord.service.js';
/** Walk up from cwd to find the monorepo root (has pnpm-workspace.yaml). */
function findMonorepoRoot(start: string): string {
let dir = start;
for (let i = 0; i < 5; i++) {
try {
fs.accessSync(path.join(dir, 'pnpm-workspace.yaml'));
return dir;
} catch {
const parent = path.dirname(dir);
if (parent === dir) break;
dir = parent;
}
}
return start;
}
/** Only paths under these roots are allowed for coord queries. */
const WORKSPACE_ROOT = process.env['MOSAIC_WORKSPACE_ROOT'] ?? findMonorepoRoot(process.cwd());
const ALLOWED_ROOTS = [process.cwd(), WORKSPACE_ROOT];
function resolveAndValidatePath(raw: string | undefined): string {
const resolved = path.resolve(raw ?? process.cwd());
const isAllowed = ALLOWED_ROOTS.some(
(root) => resolved === root || resolved.startsWith(`${root}/`),
);
if (!isAllowed) {
throw new BadRequestException('projectPath is outside the allowed workspace');
}
return resolved;
}
/**
* File-based coord endpoints for agent tool consumption.
* DB-backed mission CRUD has moved to MissionsController at /api/missions.
*/
@Controller('api/coord')
@UseGuards(AuthGuard)
export class CoordController {
constructor(@Inject(CoordService) private readonly coordService: CoordService) {}
@Get('status')
async missionStatus(@Query('projectPath') projectPath?: string) {
const resolvedPath = resolveAndValidatePath(projectPath);
const status = await this.coordService.getMissionStatus(resolvedPath);
if (!status) throw new NotFoundException('No active coord mission found');
return status;
}
@Get('tasks')
async listTasks(@Query('projectPath') projectPath?: string) {
const resolvedPath = resolveAndValidatePath(projectPath);
return this.coordService.listTasks(resolvedPath);
}
@Get('tasks/:taskId')
async taskStatus(@Param('taskId') taskId: string, @Query('projectPath') projectPath?: string) {
const resolvedPath = resolveAndValidatePath(projectPath);
const detail = await this.coordService.getTaskStatus(resolvedPath, taskId);
if (!detail) throw new NotFoundException(`Task ${taskId} not found in coord mission`);
return detail;
}
}
-90
View File
@@ -1,90 +0,0 @@
// ── File-based coord DTOs (legacy file-system backed) ──
export interface CoordMissionStatusDto {
mission: {
id: string;
name: string;
status: string;
projectPath: string;
};
milestones: {
total: number;
completed: number;
current?: {
id: string;
name: string;
status: string;
};
};
tasks: {
total: number;
done: number;
inProgress: number;
pending: number;
blocked: number;
cancelled: number;
};
nextTaskId?: string;
activeSession?: {
sessionId: string;
runtime: string;
startedAt: string;
};
}
export interface CoordTaskDetailDto {
missionId: string;
task: {
id: string;
title: string;
status: string;
milestone?: string;
pr?: string;
notes?: string;
};
isNextTask: boolean;
activeSession?: {
sessionId: string;
runtime: string;
startedAt: string;
};
}
// ── DB-backed coord DTOs ──
export interface CreateDbMissionDto {
name: string;
description?: string;
projectId?: string;
phase?: string;
milestones?: Record<string, unknown>[];
config?: Record<string, unknown>;
status?: 'planning' | 'active' | 'paused' | 'completed' | 'failed';
}
export interface UpdateDbMissionDto {
name?: string;
description?: string;
projectId?: string;
phase?: string;
milestones?: Record<string, unknown>[];
config?: Record<string, unknown>;
status?: 'planning' | 'active' | 'paused' | 'completed' | 'failed';
}
export interface CreateMissionTaskDto {
missionId: string;
taskId?: string;
status?: 'not-started' | 'in-progress' | 'blocked' | 'done' | 'cancelled';
description?: string;
notes?: string;
pr?: string;
}
export interface UpdateMissionTaskDto {
taskId?: string;
status?: 'not-started' | 'in-progress' | 'blocked' | 'done' | 'cancelled';
description?: string;
notes?: string;
pr?: string;
}
-10
View File
@@ -1,10 +0,0 @@
import { Module } from '@nestjs/common';
import { CoordService } from './coord.service.js';
import { CoordController } from './coord.controller.js';
@Module({
providers: [CoordService],
controllers: [CoordController],
exports: [CoordService],
})
export class CoordModule {}
-77
View File
@@ -1,77 +0,0 @@
import { Injectable, Logger } from '@nestjs/common';
import {
loadMission,
getMissionStatus,
getTaskStatus,
parseTasksFile,
type Mission,
type MissionStatusSummary,
type MissionTask,
type TaskDetail,
} from '@mosaic/coord';
import { promises as fs } from 'node:fs';
import path from 'node:path';
/**
* File-based coord operations for agent tool consumption.
* DB-backed mission CRUD is handled directly by MissionsController via Brain repos.
*/
@Injectable()
export class CoordService {
private readonly logger = new Logger(CoordService.name);
async loadMission(projectPath: string): Promise<Mission | null> {
try {
return await loadMission(projectPath);
} catch (err) {
this.logger.debug(
`No coord mission at ${projectPath}: ${err instanceof Error ? err.message : String(err)}`,
);
return null;
}
}
async getMissionStatus(projectPath: string): Promise<MissionStatusSummary | null> {
const mission = await this.loadMission(projectPath);
if (!mission) return null;
try {
return await getMissionStatus(mission);
} catch (err) {
this.logger.error(
`Failed to get mission status: ${err instanceof Error ? err.message : String(err)}`,
);
return null;
}
}
async getTaskStatus(projectPath: string, taskId: string): Promise<TaskDetail | null> {
const mission = await this.loadMission(projectPath);
if (!mission) return null;
try {
return await getTaskStatus(mission, taskId);
} catch (err) {
this.logger.error(
`Failed to get task status for ${taskId}: ${err instanceof Error ? err.message : String(err)}`,
);
return null;
}
}
async listTasks(projectPath: string): Promise<MissionTask[]> {
const mission = await this.loadMission(projectPath);
if (!mission) return [];
const tasksFile = path.isAbsolute(mission.tasksFile)
? mission.tasksFile
: path.join(mission.projectPath, mission.tasksFile);
try {
const content = await fs.readFile(tasksFile, 'utf8');
return parseTasksFile(content);
} catch {
return [];
}
}
}
@@ -1,28 +0,0 @@
import { Global, Inject, Module, type OnApplicationShutdown } from '@nestjs/common';
import { createDb, type Db, type DbHandle } from '@mosaic/db';
export const DB_HANDLE = 'DB_HANDLE';
export const DB = 'DB';
@Global()
@Module({
providers: [
{
provide: DB_HANDLE,
useFactory: (): DbHandle => createDb(),
},
{
provide: DB,
useFactory: (handle: DbHandle): Db => handle.db,
inject: [DB_HANDLE],
},
],
exports: [DB],
})
export class DatabaseModule implements OnApplicationShutdown {
constructor(@Inject(DB_HANDLE) private readonly handle: DbHandle) {}
async onApplicationShutdown(): Promise<void> {
await this.handle.close();
}
}
-31
View File
@@ -1,31 +0,0 @@
import { Module, type OnApplicationShutdown, Inject } from '@nestjs/common';
import { createQueue, type QueueHandle } from '@mosaic/queue';
import { SessionGCService } from './session-gc.service.js';
import { REDIS } from './gc.tokens.js';
const GC_QUEUE_HANDLE = 'GC_QUEUE_HANDLE';
@Module({
providers: [
{
provide: GC_QUEUE_HANDLE,
useFactory: (): QueueHandle => {
return createQueue();
},
},
{
provide: REDIS,
useFactory: (handle: QueueHandle) => handle.redis,
inject: [GC_QUEUE_HANDLE],
},
SessionGCService,
],
exports: [SessionGCService],
})
export class GCModule implements OnApplicationShutdown {
constructor(@Inject(GC_QUEUE_HANDLE) private readonly handle: QueueHandle) {}
async onApplicationShutdown(): Promise<void> {
await this.handle.close().catch(() => {});
}
}
-1
View File
@@ -1 +0,0 @@
export const REDIS = 'REDIS';
@@ -1,97 +0,0 @@
import { describe, it, expect, vi, beforeEach } from 'vitest';
import { Logger } from '@nestjs/common';
import type { QueueHandle } from '@mosaic/queue';
import type { LogService } from '@mosaic/log';
import { SessionGCService } from './session-gc.service.js';
type MockRedis = {
keys: ReturnType<typeof vi.fn>;
del: ReturnType<typeof vi.fn>;
};
describe('SessionGCService', () => {
let service: SessionGCService;
let mockRedis: MockRedis;
let mockLogService: { logs: { promoteToWarm: ReturnType<typeof vi.fn> } };
beforeEach(() => {
mockRedis = {
keys: vi.fn().mockResolvedValue([]),
del: vi.fn().mockResolvedValue(0),
};
mockLogService = {
logs: {
promoteToWarm: vi.fn().mockResolvedValue(0),
},
};
// Suppress logger output in tests
vi.spyOn(Logger.prototype, 'log').mockImplementation(() => {});
service = new SessionGCService(
mockRedis as unknown as QueueHandle['redis'],
mockLogService as unknown as LogService,
);
});
it('collect() deletes Valkey keys for session', async () => {
mockRedis.keys.mockResolvedValue(['mosaic:session:abc:system', 'mosaic:session:abc:foo']);
const result = await service.collect('abc');
expect(mockRedis.del).toHaveBeenCalledWith(
'mosaic:session:abc:system',
'mosaic:session:abc:foo',
);
expect(result.cleaned.valkeyKeys).toBe(2);
});
it('collect() with no keys returns empty cleaned valkeyKeys', async () => {
mockRedis.keys.mockResolvedValue([]);
const result = await service.collect('abc');
expect(result.cleaned.valkeyKeys).toBeUndefined();
});
it('collect() returns sessionId in result', async () => {
const result = await service.collect('test-session-id');
expect(result.sessionId).toBe('test-session-id');
});
it('fullCollect() deletes all session keys', async () => {
mockRedis.keys.mockResolvedValue(['mosaic:session:abc:system', 'mosaic:session:xyz:foo']);
const result = await service.fullCollect();
expect(mockRedis.del).toHaveBeenCalled();
expect(result.valkeyKeys).toBe(2);
});
it('fullCollect() with no keys returns 0 valkeyKeys', async () => {
mockRedis.keys.mockResolvedValue([]);
const result = await service.fullCollect();
expect(result.valkeyKeys).toBe(0);
expect(mockRedis.del).not.toHaveBeenCalled();
});
it('fullCollect() returns duration', async () => {
const result = await service.fullCollect();
expect(result.duration).toBeGreaterThanOrEqual(0);
});
it('sweepOrphans() extracts unique session IDs and collects them', async () => {
mockRedis.keys.mockResolvedValue([
'mosaic:session:abc:system',
'mosaic:session:abc:messages',
'mosaic:session:xyz:system',
]);
mockRedis.del.mockResolvedValue(1);
const result = await service.sweepOrphans();
expect(result.orphanedSessions).toBeGreaterThanOrEqual(0);
expect(result.duration).toBeGreaterThanOrEqual(0);
});
it('sweepOrphans() returns empty when no session keys', async () => {
mockRedis.keys.mockResolvedValue([]);
const result = await service.sweepOrphans();
expect(result.orphanedSessions).toBe(0);
expect(result.totalCleaned).toHaveLength(0);
});
});
-139
View File
@@ -1,139 +0,0 @@
import { Inject, Injectable, Logger, type OnModuleInit } from '@nestjs/common';
import type { QueueHandle } from '@mosaic/queue';
import type { LogService } from '@mosaic/log';
import { LOG_SERVICE } from '../log/log.tokens.js';
import { REDIS } from './gc.tokens.js';
export interface GCResult {
sessionId: string;
cleaned: {
valkeyKeys?: number;
logsDemoted?: number;
tempFilesRemoved?: number;
};
}
export interface GCSweepResult {
orphanedSessions: number;
totalCleaned: GCResult[];
duration: number;
}
export interface FullGCResult {
valkeyKeys: number;
logsDemoted: number;
jobsPurged: number;
tempFilesRemoved: number;
duration: number;
}
@Injectable()
export class SessionGCService implements OnModuleInit {
private readonly logger = new Logger(SessionGCService.name);
constructor(
@Inject(REDIS) private readonly redis: QueueHandle['redis'],
@Inject(LOG_SERVICE) private readonly logService: LogService,
) {}
async onModuleInit(): Promise<void> {
this.logger.log('Running full GC on cold start...');
const result = await this.fullCollect();
this.logger.log(
`Full GC complete: ${result.valkeyKeys} Valkey keys, ` +
`${result.logsDemoted} logs demoted, ` +
`${result.jobsPurged} jobs purged, ` +
`${result.tempFilesRemoved} temp dirs removed ` +
`(${result.duration}ms)`,
);
}
/**
* Immediate cleanup for a single session (call from destroySession).
*/
async collect(sessionId: string): Promise<GCResult> {
const result: GCResult = { sessionId, cleaned: {} };
// 1. Valkey: delete all session-scoped keys
const pattern = `mosaic:session:${sessionId}:*`;
const valkeyKeys = await this.redis.keys(pattern);
if (valkeyKeys.length > 0) {
await this.redis.del(...valkeyKeys);
result.cleaned.valkeyKeys = valkeyKeys.length;
}
// 2. PG: demote hot-tier agent_logs for this session to warm
const cutoff = new Date(); // demote all hot logs for this session
const logsDemoted = await this.logService.logs.promoteToWarm(cutoff);
if (logsDemoted > 0) {
result.cleaned.logsDemoted = logsDemoted;
}
return result;
}
/**
* Sweep GC — find orphaned artifacts from dead sessions.
* User-scoped when userId provided; system-wide when null (admin).
*/
async sweepOrphans(_userId?: string): Promise<GCSweepResult> {
const start = Date.now();
const cleaned: GCResult[] = [];
// 1. Find all session-scoped Valkey keys
const allSessionKeys = await this.redis.keys('mosaic:session:*');
// Extract unique session IDs from keys
const sessionIds = new Set<string>();
for (const key of allSessionKeys) {
const match = key.match(/^mosaic:session:([^:]+):/);
if (match) sessionIds.add(match[1]!);
}
// 2. For each session ID, collect stale keys
for (const sessionId of sessionIds) {
const gcResult = await this.collect(sessionId);
if (Object.keys(gcResult.cleaned).length > 0) {
cleaned.push(gcResult);
}
}
return {
orphanedSessions: cleaned.length,
totalCleaned: cleaned,
duration: Date.now() - start,
};
}
/**
* Full GC — aggressive collection for cold start.
* Assumes no sessions survived the restart.
*/
async fullCollect(): Promise<FullGCResult> {
const start = Date.now();
// 1. Valkey: delete ALL session-scoped keys
const sessionKeys = await this.redis.keys('mosaic:session:*');
if (sessionKeys.length > 0) {
await this.redis.del(...sessionKeys);
}
// 2. NOTE: channel keys are NOT collected on cold start
// (discord/telegram plugins may reconnect and resume)
// 3. PG: demote stale hot-tier logs older than 24h to warm
const hotCutoff = new Date(Date.now() - 24 * 60 * 60 * 1000);
const logsDemoted = await this.logService.logs.promoteToWarm(hotCutoff);
// 4. No summarization job purge API available yet
const jobsPurged = 0;
return {
valkeyKeys: sessionKeys.length,
logsDemoted,
jobsPurged,
tempFilesRemoved: 0,
duration: Date.now() - start,
};
}
}
@@ -1,9 +0,0 @@
import { Controller, Get } from '@nestjs/common';
@Controller('health')
export class HealthController {
@Get()
check(): { status: string } {
return { status: 'ok' };
}
}
-63
View File
@@ -1,63 +0,0 @@
import {
Inject,
Injectable,
Logger,
type OnModuleInit,
type OnModuleDestroy,
} from '@nestjs/common';
import cron from 'node-cron';
import { SummarizationService } from './summarization.service.js';
import { SessionGCService } from '../gc/session-gc.service.js';
@Injectable()
export class CronService implements OnModuleInit, OnModuleDestroy {
private readonly logger = new Logger(CronService.name);
private readonly tasks: cron.ScheduledTask[] = [];
constructor(
@Inject(SummarizationService) private readonly summarization: SummarizationService,
@Inject(SessionGCService) private readonly sessionGC: SessionGCService,
) {}
onModuleInit(): void {
const summarizationSchedule = process.env['SUMMARIZATION_CRON'] ?? '0 */6 * * *'; // every 6 hours
const tierManagementSchedule = process.env['TIER_MANAGEMENT_CRON'] ?? '0 3 * * *'; // daily at 3am
const gcSchedule = process.env['SESSION_GC_CRON'] ?? '0 4 * * *'; // daily at 4am
this.tasks.push(
cron.schedule(summarizationSchedule, () => {
this.summarization.runSummarization().catch((err) => {
this.logger.error(`Scheduled summarization failed: ${err}`);
});
}),
);
this.tasks.push(
cron.schedule(tierManagementSchedule, () => {
this.summarization.runTierManagement().catch((err) => {
this.logger.error(`Scheduled tier management failed: ${err}`);
});
}),
);
this.tasks.push(
cron.schedule(gcSchedule, () => {
this.sessionGC.sweepOrphans().catch((err) => {
this.logger.error(`Session GC sweep failed: ${err}`);
});
}),
);
this.logger.log(
`Cron scheduled: summarization="${summarizationSchedule}", tier="${tierManagementSchedule}", gc="${gcSchedule}"`,
);
}
onModuleDestroy(): void {
for (const task of this.tasks) {
task.stop();
}
this.tasks.length = 0;
this.logger.log('Cron tasks stopped');
}
}
-62
View File
@@ -1,62 +0,0 @@
import { Body, Controller, Get, Inject, Param, Post, Query, UseGuards } from '@nestjs/common';
import type { LogService } from '@mosaic/log';
import { LOG_SERVICE } from './log.tokens.js';
import { AuthGuard } from '../auth/auth.guard.js';
import type { IngestLogDto, QueryLogsDto } from './log.dto.js';
@Controller('api/logs')
@UseGuards(AuthGuard)
export class LogController {
constructor(@Inject(LOG_SERVICE) private readonly logService: LogService) {}
@Post()
async ingest(@Query('userId') userId: string, @Body() dto: IngestLogDto) {
return this.logService.logs.ingest({
sessionId: dto.sessionId,
userId,
level: dto.level,
category: dto.category,
content: dto.content,
metadata: dto.metadata,
});
}
@Post('batch')
async ingestBatch(@Query('userId') userId: string, @Body() dtos: IngestLogDto[]) {
const entries = dtos.map((dto) => ({
sessionId: dto.sessionId,
userId,
level: dto.level as 'debug' | 'info' | 'warn' | 'error' | undefined,
category: dto.category as
| 'decision'
| 'tool_use'
| 'learning'
| 'error'
| 'general'
| undefined,
content: dto.content,
metadata: dto.metadata,
}));
return this.logService.logs.ingestBatch(entries);
}
@Get()
async query(@Query('userId') userId: string, @Query() params: QueryLogsDto) {
return this.logService.logs.query({
userId,
sessionId: params.sessionId,
level: params.level,
category: params.category,
tier: params.tier,
since: params.since ? new Date(params.since) : undefined,
until: params.until ? new Date(params.until) : undefined,
limit: params.limit ? Number(params.limit) : undefined,
offset: params.offset ? Number(params.offset) : undefined,
});
}
@Get(':id')
async findOne(@Param('id') id: string) {
return this.logService.logs.findById(id);
}
}
-18
View File
@@ -1,18 +0,0 @@
export interface IngestLogDto {
sessionId: string;
level?: 'debug' | 'info' | 'warn' | 'error';
category?: 'decision' | 'tool_use' | 'learning' | 'error' | 'general';
content: string;
metadata?: Record<string, unknown>;
}
export interface QueryLogsDto {
sessionId?: string;
level?: 'debug' | 'info' | 'warn' | 'error';
category?: 'decision' | 'tool_use' | 'learning' | 'error' | 'general';
tier?: 'hot' | 'warm' | 'cold';
since?: string;
until?: string;
limit?: string;
offset?: string;
}
-26
View File
@@ -1,26 +0,0 @@
import { Global, Module } from '@nestjs/common';
import { createLogService, type LogService } from '@mosaic/log';
import type { Db } from '@mosaic/db';
import { DB } from '../database/database.module.js';
import { LOG_SERVICE } from './log.tokens.js';
import { LogController } from './log.controller.js';
import { SummarizationService } from './summarization.service.js';
import { CronService } from './cron.service.js';
import { GCModule } from '../gc/gc.module.js';
@Global()
@Module({
imports: [GCModule],
providers: [
{
provide: LOG_SERVICE,
useFactory: (db: Db): LogService => createLogService(db),
inject: [DB],
},
SummarizationService,
CronService,
],
controllers: [LogController],
exports: [LOG_SERVICE, SummarizationService],
})
export class LogModule {}
-1
View File
@@ -1 +0,0 @@
export const LOG_SERVICE = 'LOG_SERVICE';
@@ -1,178 +0,0 @@
import { Inject, Injectable, Logger } from '@nestjs/common';
import type { LogService } from '@mosaic/log';
import type { Memory } from '@mosaic/memory';
import { LOG_SERVICE } from './log.tokens.js';
import { MEMORY } from '../memory/memory.tokens.js';
import { EmbeddingService } from '../memory/embedding.service.js';
import type { Db } from '@mosaic/db';
import { sql, summarizationJobs } from '@mosaic/db';
import { DB } from '../database/database.module.js';
const SUMMARIZATION_PROMPT = `You are a knowledge extraction assistant. Given the following agent interaction logs, extract the key decisions, learnings, and patterns. Output a concise summary (2-4 sentences) that captures the most important information for future reference. Focus on actionable insights, not raw events.
Logs:
{logs}
Summary:`;
interface ChatCompletion {
choices: Array<{ message: { content: string } }>;
}
@Injectable()
export class SummarizationService {
private readonly logger = new Logger(SummarizationService.name);
private readonly apiKey: string | undefined;
private readonly baseUrl: string;
private readonly model: string;
constructor(
@Inject(LOG_SERVICE) private readonly logService: LogService,
@Inject(MEMORY) private readonly memory: Memory,
@Inject(EmbeddingService) private readonly embeddings: EmbeddingService,
@Inject(DB) private readonly db: Db,
) {
this.apiKey = process.env['OPENAI_API_KEY'];
this.baseUrl = process.env['SUMMARIZATION_API_URL'] ?? 'https://api.openai.com/v1';
this.model = process.env['SUMMARIZATION_MODEL'] ?? 'gpt-4o-mini';
}
/**
* Run one summarization cycle:
* 1. Find hot logs older than 24h with decision/learning/tool_use categories
* 2. Group by session
* 3. Summarize each group via cheap LLM
* 4. Store as insights with embeddings
* 5. Transition processed logs to warm tier
*/
async runSummarization(): Promise<{ logsProcessed: number; insightsCreated: number }> {
const cutoff = new Date(Date.now() - 24 * 60 * 60 * 1000); // 24h ago
// Create job record
const [job] = await this.db
.insert(summarizationJobs)
.values({ status: 'running', startedAt: new Date() })
.returning();
try {
const logs = await this.logService.logs.getLogsForSummarization(cutoff, 200);
if (logs.length === 0) {
await this.db
.update(summarizationJobs)
.set({ status: 'completed', completedAt: new Date() })
.where(sql`id = ${job!.id}`);
return { logsProcessed: 0, insightsCreated: 0 };
}
// Group logs by session
const bySession = new Map<string, typeof logs>();
for (const log of logs) {
const group = bySession.get(log.sessionId) ?? [];
group.push(log);
bySession.set(log.sessionId, group);
}
let insightsCreated = 0;
for (const [sessionId, sessionLogs] of bySession) {
const userId = sessionLogs[0]?.userId;
if (!userId) continue;
const logsText = sessionLogs.map((l) => `[${l.category}] ${l.content}`).join('\n');
const summary = await this.summarize(logsText);
if (!summary) continue;
const embedding = this.embeddings.available
? await this.embeddings.embed(summary)
: undefined;
await this.memory.insights.create({
userId,
content: summary,
embedding: embedding ?? null,
source: 'summarization',
category: 'learning',
metadata: { sessionId, logCount: sessionLogs.length },
});
insightsCreated++;
}
// Transition processed logs to warm
await this.logService.logs.promoteToWarm(cutoff);
await this.db
.update(summarizationJobs)
.set({
status: 'completed',
logsProcessed: logs.length,
insightsCreated,
completedAt: new Date(),
})
.where(sql`id = ${job!.id}`);
this.logger.log(`Summarization complete: ${logs.length} logs → ${insightsCreated} insights`);
return { logsProcessed: logs.length, insightsCreated };
} catch (error) {
const message = error instanceof Error ? error.message : String(error);
await this.db
.update(summarizationJobs)
.set({ status: 'failed', errorMessage: message, completedAt: new Date() })
.where(sql`id = ${job!.id}`);
this.logger.error(`Summarization failed: ${message}`);
throw error;
}
}
/**
* Run tier management:
* - Warm logs older than 30 days → cold
* - Cold logs older than 90 days → purged
* - Decay old insight relevance scores
*/
async runTierManagement(): Promise<void> {
const warmCutoff = new Date(Date.now() - 30 * 24 * 60 * 60 * 1000);
const coldCutoff = new Date(Date.now() - 90 * 24 * 60 * 60 * 1000);
const decayCutoff = new Date(Date.now() - 14 * 24 * 60 * 60 * 1000);
const promoted = await this.logService.logs.promoteToCold(warmCutoff);
const purged = await this.logService.logs.purge(coldCutoff);
const decayed = await this.memory.insights.decayOldInsights(decayCutoff);
this.logger.log(
`Tier management: ${promoted} logs→cold, ${purged} purged, ${decayed} insights decayed`,
);
}
private async summarize(logsText: string): Promise<string | null> {
if (!this.apiKey) {
this.logger.warn('No API key configured — skipping summarization');
return null;
}
const prompt = SUMMARIZATION_PROMPT.replace('{logs}', logsText);
const response = await fetch(`${this.baseUrl}/chat/completions`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
Authorization: `Bearer ${this.apiKey}`,
},
body: JSON.stringify({
model: this.model,
messages: [{ role: 'user', content: prompt }],
max_tokens: 300,
temperature: 0.3,
}),
});
if (!response.ok) {
const body = await response.text();
this.logger.error(`Summarization API error: ${response.status} ${body}`);
return null;
}
const json = (await response.json()) as ChatCompletion;
return json.choices[0]?.message.content ?? null;
}
}

Some files were not shown because too many files have changed in this diff Show More