All six HTTP-status error arms captured the provider's response body in a mktemp file and then discarded it unread, so a caller got a status code and no cause. The review-submit arm additionally hardcoded "(#865: no durable review created)" onto EVERY non-2xx — including 401/403/404/409/422, where the server had given a definite, correct, machine-readable refusal that is categorically not #865 (the silent-no-op defect). Measured case: `pr-review.sh -n 1001 -a approve` reported Error: Gitea review submit failed with HTTP 422 (#865: no durable review created) when the real cause was that the acting credential authored the PR and Gitea correctly refuses self-approval. Gitea said so plainly in the discarded body. This matters beyond error aesthetics: withholding the cause pushes the caller toward re-issuing the request by hand to see what the server says, and a diagnostic that uses the real verb against the real object is not a diagnostic, it is the operation. Changes: - New gitea_error_detail() renders the provider's own explanation (JSON message/error/errors, else the first body line), truncated to 300 chars. It always returns 0 so it can never abort a caller under `set -e`, and prints nothing when there is nothing to add. - All six arms append it. The review-submit arm no longer cites #865; that reference now appears only on the check that actually diagnoses it — a 2xx yielding no id, immediately below. No control flow changes: every arm still returns 1 and still fails closed. Test hermeticity (required, or the new cases cannot run on any agent seat): get_gitea_token()'s per-agent identity step reads `git config --get mosaic.gitIdentity`, which on a provisioned seat is set GLOBALLY and leaks into the harness's fresh `git init` repo. It then returns a REAL per-slot token from $HOME without ever consulting MOSAIC_CREDENTIALS_FILE, so the fixture placeholder was silently ignored and the suite ran against production credentials, failing 401. The harness now pins an empty repo-local mosaic.gitIdentity (an empty local value shadows the global and reads back empty at rc=0) and runs under a sandboxed HOME. Filed separately: the per-slot token store has no override hook, so no harness can sandbox it. Verification matrix (all three run locally): - base suite + hermeticity fix, unmodified wrapper -> PASS (fix is sufficient and breaks no existing case) - new cases 19/20, unmodified wrapper -> FAIL at case 19, reproducing the exact misattribution above (negative control) - new cases 19/20, fixed wrapper -> PASS Cases 19 (422 JSON refusal) and 20 (502 HTML body) assert the status is present, the provider's reason is present, and "#865: no durable review created" is ABSENT. shellcheck is not installed on this host, so the lint gate is unverified locally. Co-authored-by: mos-dt-0 <[email protected]>
@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="[email protected]"
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.