Capture user corrections as high-importance BrainLayer entries. Use when user says no/wrong/stop, repeats instructions, expresses frustration, or during session mining. Triggers on: user correction, 'I told you', 'not that', frustration signal.
Create, edit, audit, and evaluate golem-powers skills. Use for new skills, structural skill edits, workflow/adapter changes, pre-deploy validation, skill evals, benchmarks, live A/B tests, session JSONL mining, batch miners, and handoff digests. Triggers: create skill, edit skill, audit skill, validate skill, skill eval, live eval, mine session. NOT for invoking existing skills or convergence weaving.
Brain Drive filing discipline — where every artifact goes + how to name it. Use WHENEVER touching Google Drive / Brain Drive: uploading, creating folders, saving research prompts/results, audits, plans, transcripts, dashboards, or when about to leave a durable artifact in docs.local/. Teaches the numbered folder model (01_STANDARDS / 02_GROUNDING / 03_RESEARCH / 04_INGEST / 06_ARCHIVE), date-prefixed naming, and the rule: FILE durable artifacts in the right Drive folder — docs.local/ is cache-only. NOT for querying Drive via Gemini (use /braindrive) or web research (use /gemini-research); for >100KB heavy archival defer to /google-drive-archive.
Launch agents in any repo via repoGolem launchers ({name}Claude, {name}Codex, {name}Cursor). Unified flags: -s skip, -c continue, -m model, -p headless, -w worktree cwd. 40 projects registered. Triggers on: spawn agent, launcher, repoGolem, brainlayerClaude, flags.
Orchestrator-only convergence workflow: mine recent Claude/Codex JSONLs, weave cited findings into an action ledger, route every finding to a disposition, then red-team facts against raw logs. Triggers: weave, weave now, run weave, session weave, convergence weave. Use only when fleet is quiet; NOT for single-session mining or web research.
Pre-R0 sprint gate that diffs implementation vs SOTA research output verbatim. Surfaces cited counter-examples and architectural mismatches before sprint hooks fire. Triggers: 'before R0', 'architectural audit', 'verify against research'. NOT for per-PR review or post-merge.
Full code review lifecycle: requesting reviews (CodeRabbit, Greptile, Bugbot, GitHub PR comments) and receiving feedback (classify issues, implement fixes, push back on wrong suggestions). Use when: creating a PR review, reading review comments, handling reviewer feedback, fixing review items, or deciding whether to accept or reject a suggestion. NOT for: running tests directly or CI/CD pipeline issues (use relevant repo tools).
Use when reviewing uncommitted changes, preparing PRs, requesting or receiving code review, handling CodeRabbit/Greptile/Bugbot/GitHub PR comments, checking security/secrets/a11y/code quality, or deciding whether to accept or reject reviewer feedback. Runs AI review via CLI and covers review triage, false-positive pushback, red/blue team profiles, PR-ready gates. NOT for: runtime debugging or test execution.