Step 5 of v2.0 plan + critical pipeline fix. Stop hook (hooks/scripts/stop-context-monitor.mjs): - Estimates context usage from transcript size (chars/3.5 / window_size) - At ≥70%, spawns handoff-pipeline.mjs --auto --no-push synchronously - Reads context_window_size from payload (supports 1M windows) - Lock file at <transcript_dir>/.handoff-lock-<session_id> - Gracefully handles missing CLAUDE_PLUGIN_ROOT, missing transcript Pipeline fix (scripts/handoff-pipeline.mjs): - REMOVED `git add -A` (CLAUDE.md anti-pattern: scoops up unrelated WIP) - Now stages ONLY artifact + REMEMBER.md/TODO.md if present - New regression test 'pipeline never stages unrelated dirty files' Tests: 7 stop-hook tests use stub pipeline (no real git operations); 11 pipeline tests including new regression for explicit staging. |
||
|---|---|---|
| .. | ||
| ai-psychosis | ||
| config-audit | ||
| graceful-handoff | ||
| linkedin-thought-leadership | ||
| llm-security | ||
| ms-ai-architect | ||
| okr | ||
| ultra-cc-architect | ||
| ultraplan-local | ||