84d802cafe
The per-agent identity resolution added to get_gitea_token() (Patch 2 / #873) resolves an explicit identity (MOSAIC_GIT_IDENTITY env, or git config mosaic.gitIdentity) and returns that identity's per-slot Gitea token when present. Gap: when the identity resolves but its per-slot token file is ABSENT for a recognized Gitea host, the function fell through to the shared/default credential-loader token instead of failing. API tooling (pr-create.sh, issue-create.sh, pr-review.sh) then silently posted the PR/issue/review as the WRONG agent — e.g. a reviewer seat's Gate-16 review getting attributed to the shared/default identity — corrupting author≠reviewer separation while also masking the missing-token misconfiguration. This surfaced most often on the tea-stale API-fallback path, which is exactly when tooling leans on get_gitea_token. Fix: in the step-0 explicit-identity branch, when the per-slot token for that identity is absent on a recognized Gitea host, print a stderr diagnostic naming the identity, its source (env vs git config), the host, and the expected token path, then return 1 instead of falling through. All three callers already `return 1` on a nonzero get_gitea_token, so fail-loud propagates with zero caller edits. Scope (deliberate, minimal blast radius): explicit-identity-only. Plain `git config user.name` is not an identity trigger — only MOSAIC_GIT_IDENTITY / git config mosaic.gitIdentity count, so ordinary shared/human repo usage is unaffected. Recognized-Gitea-hosts-only: unrecognized hosts have no per-slot token scheme, so identity-set + unknown-host still falls through unchanged (no fail-loud). The MOSAIC_STRICT_IDENTITY opt-in discussed as a possible future extension (gating the full `... > git username` chain) is deliberately NOT part of this patch — deferred per spec as a later, separate proposal. Red-first: stashed the source fix, ran the extended test-gitea-token-identity.sh — 16 assertions failed exactly as expected (shared token leaked, no fail-loud diagnostic). Restored the fix — all green, full test:framework-shell chain passes. Backward compat verified: the no-identity-requested case still returns the shared token unchanged. Extends test-gitea-token-identity.sh: the no-per-slot-token case (both identity sources: env and git config) on a recognized host now asserts nonzero return + empty stdout + stderr diagnostic naming identity/source/host/expected path, instead of asserting a shared-token fallback. Adds a same-identity cross-host case (token exists for one host, absent for another → fail-loud on the host lacking it, no cross-host token leak) and a scope-containment case (identity set + unrecognized host → existing fall-through behavior unchanged, fail-loud diagnostic does not fire). Gates: shellcheck clean on both changed files, sanitization gate (verify-sanitized.sh) passes, full test:framework-shell chain green. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0158NZqN2n2ymKFeJAZ4GUCb
@mosaicstack/mosaic
CLI package for the Mosaic self-hosted AI agent platform.
Usage
mosaic wizard # First-run setup wizard
mosaic gateway install # Install the gateway daemon
mosaic config show # View current configuration
mosaic config hooks list # Manage Claude hooks
Headless / CI Installation
Set MOSAIC_ASSUME_YES=1 (or ensure stdin is not a TTY) to skip all interactive prompts. The following environment variables control the install:
Gateway configuration (mosaic gateway install)
| Variable | Default | Required |
|---|---|---|
MOSAIC_STORAGE_TIER |
local |
No |
MOSAIC_GATEWAY_PORT |
14242 |
No |
MOSAIC_DATABASE_URL |
(none) | Yes if tier=team |
MOSAIC_VALKEY_URL |
(none) | Yes if tier=team |
MOSAIC_ANTHROPIC_API_KEY |
(none) | No |
MOSAIC_CORS_ORIGIN |
http://localhost:3000 |
No |
Admin user bootstrap
| Variable | Default | Required |
|---|---|---|
MOSAIC_ADMIN_NAME |
(none) | Yes (headless) |
MOSAIC_ADMIN_EMAIL |
(none) | Yes (headless) |
MOSAIC_ADMIN_PASSWORD |
(none) | Yes (headless) |
MOSAIC_ADMIN_PASSWORD must be at least 8 characters. In headless mode a missing or too-short password causes a non-zero exit.
Example: Docker / CI install
export MOSAIC_ASSUME_YES=1
export MOSAIC_ADMIN_NAME="Admin"
export MOSAIC_ADMIN_EMAIL="admin@example.com"
export MOSAIC_ADMIN_PASSWORD="securepass123"
mosaic gateway install
Runtime launchers
mosaic claude # Launch Claude Code with Mosaic injection
mosaic yolo claude # …with --dangerously-skip-permissions
mosaic codex | opencode | pi
mosaic claudex (EXPERIMENTAL)
Runs GPT models inside the Claude Code harness by pointing Claude Code at a
local claude-code-proxy that
translates the Anthropic Messages API to a ChatGPT-subscription (Codex OAuth)
backend. This is not Anthropic Claude — model behavior, tool use, and output
quality may differ. Intended for evaluation, not production delivery.
mosaic claudex # launch (prompts through the proxy readiness gate)
mosaic yolo claudex # …with --dangerously-skip-permissions
mosaic claudex --print "hello" # trailing args are forwarded to Claude Code
Prerequisite: the claude-code-proxy binary must be installed and
authenticated (claude-code-proxy codex auth …). mosaic claudex runs a
preflight that verifies the binary, the OAuth state (triggering a device re-auth
if needed), and a trusted local listener before launching; it fails closed
if the proxy cannot be brought up with a verified identity.
Isolation (never touches your real Claude state). claudex always launches
against an isolated CLAUDE_CONFIG_DIR (default ~/.config/mosaic/claudex/home).
The ambient CLAUDE_CONFIG_DIR is deliberately ignored, and a guard proves the
resolved dir can never be — or live under — the real ~/.claude. A claudex
session therefore cannot mutate your normal Claude Code config.
No token leakage. claudex never reads the proxy's credential file. Claude
Code is handed only ANTHROPIC_AUTH_TOKEN=unused pointed at the loopback proxy;
the entire credential-bearing env family (ANTHROPIC_*, AWS_*, GOOGLE_CLOUD_*,
GOOGLE_APPLICATION_CREDENTIALS, *_TOKEN, *_KEY, *_SECRET, …) is stripped
from the composed environment. The Bedrock/Vertex routing switches
(CLAUDE_CODE_USE_BEDROCK, CLAUDE_CODE_USE_VERTEX, and the _SKIP_*_AUTH
pair) are force-removed regardless of value — otherwise their mere presence
would route Claude Code to the real Anthropic API via AWS/GCP and bypass the
proxy. The proxy holds the real OAuth credential.
Model tiers (override via env).
| Tier | Env var | Default |
|---|---|---|
| primary (opus/sonnet) | ANTHROPIC_MODEL |
gpt-5.6-sol |
| small/fast (haiku) | ANTHROPIC_SMALL_FAST_MODEL |
gpt-5.6-luna |
Operator-provided values win over the defaults. Additional overrides:
MOSAIC_CLAUDEX_CONFIG_DIR (isolated config dir), ANTHROPIC_BASE_URL (proxy
endpoint).
Hooks management
After running mosaic wizard, Claude hooks are installed in ~/.claude/hooks-config.json.
mosaic config hooks list # Show all hooks and enabled/disabled status
mosaic config hooks disable PostToolUse # Disable a hook (reversible)
mosaic config hooks enable PostToolUse # Re-enable a disabled hook
Set CLAUDE_HOME to override the default ~/.claude directory.