Lap Tran
bc7daf8590
- **Restore 82 completed tasks** from tasks/complete/ back to tasks/ top level (all <7 days old per the cleanup policy; premature bulk archive was fixed). - **Dashboard: fix scroll-reset on auto-refresh** — renderBoard rebuilds the board via innerHTML every 2s, destroying each column-body's scrollTop. Now snapshots column-body scrollTop + board.scrollLeft + view.scrollTop before rebuild and restores after (matched by PHASE_GROUPS index). - **Dashboard UI additions** (pre-existing unstaged work): approval section cards, transition buttons, inline artifact editor (textarea for writing missing SPEC/VERDICT/etc from the detail modal). - **Bind ornith as Implement model** — config.md: Model explicit to omlx/Ornith-1.0-35B-4bit-mlx, context window 32768. Interactive autopilot already used ornith via opencode default; now explicit. - **Fix cleanup stub** — automaton-cleanup.sh had a stale --project arg pointing at a pytest temp dir (test isolation leak). Rewired to point at ~/.automaton. - **Fix plist-isolation test** — test asserted host plist doesn't exist, but a real install creates it. Now snapshots mtime before run, asserts unchanged after (only a write during the test counts as bleed). - **New Playwright smoke test** (tests/test_dashboard_ui.py) — 2 tests: board renders tasks, column scroll survives auto-refresh tick. Verified the test fails without the scroll fix (scrollTop resets to 0). Skipped via importorskip when playwright is absent (main CI stays green). - **Clarify SI loop scope in README** — new-project onboarding section documents the framework-scoped self-improvement loop and options (leave/pause/create project loop). - **CHANGELOG** documents all changes including the known model-divergence gap (mde tasks marked complete but per-role model binding was never implemented).
17 lines
1.7 KiB
Markdown
17 lines
1.7 KiB
Markdown
# Adversarial Bug Report: Status Script
|
|
|
|
## Deep Review
|
|
The 980-line status.py is comprehensive. State machine transitions are correctly validated. Approval gates are enforced. Atomic writes prevent corruption. Error messages are clear and actionable.
|
|
|
|
## Potential Issues
|
|
1. **Audit Category 3 stub**: The git-based unauthorized modification check is stubbed. In projects that are git repos, the audit cannot detect code edits made during non-implement phases. This is a gap — an agent could edit code during research and the audit wouldn't catch it unless the artifacts reveal it.
|
|
|
|
2. **`--same-session` heuristic is weak**: The 30-minute window for session detection is a best-effort heuristic. It cannot reliably distinguish sessions across agent restarts. The spec acknowledges this, but the implementation doesn't add much beyond mtime comparison.
|
|
|
|
3. **No `--force` flag for coordinator overrides**: The multi-agent spec defines `--force` for coordinator force-transitions, but the current implementation returns "ERROR: unknown flag" for `--force`. This is expected to be added in the multi-agent-support task but a truly adversarial agent could use this gap.
|
|
|
|
4. **Symlink attack on `.state.tmp`**: An attacker with filesystem access could create a symlink at `.state.tmp` pointing to a sensitive file, causing the atomic write to overwrite it. This is a local-privilege scenario, not a remote attack.
|
|
|
|
5. **No rate limiting on `--create-task`**: A script could create thousands of task folders. Mitigation exists via kebab-case validation, but no limit on creation count.
|
|
|
|
## Verdict: PASS — the audit stub is a known gap. No logic flaws that would compromise phase enforcement. |