메뉴
SPEC-026 outcomes ledger writers/readers plus CDV-187 read-only rollup. Helpers: emit-outcome.sh (write), outcome-rates.sh (advisory rates), rollup.sh (display). User-facing entry: /metrics.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Phase-gated bug investigation → root-cause → fix → verify cycle. Enforces root-cause-before-edit, failing-test-first, holistic callsite scan, reopen/redesign gates (SPEC-029), multi-surface done matrix, and self-calibration checklist before any "done" claim. Subcommands: /debug <desc> (full), /debug patch <desc> (fast path), /debug arch <desc> (design-first → /kickoff handoff).
Brutally honest review of staged/modified files — no sugar-coating. Thin wrapper over the council engine with preset `diff-mode`: 5 specialist investigators (logic, security, compliance, quality, simplification) run in parallel, filtered at confidence 80. Blocks commit on critical or compliance findings. Optional path argument saves the review to a file.
Fact-dense rewrite of agent memory prose (tier-0 notes, digests) without losing technical substance. Companion to /memory-distill. Zero external deps.
Adversarial council tribunal engine — reality-checks claims with material evidence. Shared engine for /council and /review-and-commit (diff-mode preset). Blind investigators, evidence-or-silence rule, dual output shapes (verdict[] and finding[]), atomic verdict index at .claude/council/index.json, feedback-memory learning loop. Judge is a dedicated agent with an empty tool allowlist. See specs/core/SPEC-013-adversarial-council-tribunal.md.
Optional host SAST (Semgrep / CodeQL) feed for security review. Fail-open when tools are missing. Agent-internal + review-and-commit / council security flavor. Zero required deps.
Socratic design refinement — structured questioning that forces requirement clarification before any planning or implementation. Use before /kickoff for complex features, or standalone for early-stage ideation. Optional --grill for one-question-at-a-time interviews with recommended answers.
| name | metrics |
| description | SPEC-026 outcomes ledger writers/readers plus CDV-187 read-only rollup. Helpers: emit-outcome.sh (write), outcome-rates.sh (advisory rates), rollup.sh (display). User-facing entry: /metrics. |
Observability helpers for review-outcome routing (SPEC-026) and the all-time rollup command (CDV-187). Display and write paths are intentionally split.
skills/metrics/
├── SKILL.md (this file)
├── emit-outcome.sh SPEC-026 writer — append outcomes.jsonl
├── outcome-rates.sh SPEC-026 M5 advisory rates (read)
├── rollup.sh CDV-187 read-only multi-source rollup
└── test.sh bite-tests for emit / rates / rollup
| Path | Owner | Mode |
|---|---|---|
.claude/metrics/outcomes.jsonl | SPEC-026 / emit-outcome.sh | append-only write |
.claude/local-agent/metrics.jsonl | SPEC-019 (run.sh + emit-orch-metric.sh) | append-only write |
.claude/council/index.json | SPEC-013 | atomic index write |
/metrics + rollup.sh | CDV-187 | read-only display |
rollup.sh MUST NOT write under .claude/, call emit helpers, open council
report bodies, or re-run retro-gate.
metrics.jsonl mixes two record types:
run.sh: has outcome ∈ {success,fail,fallback}; no ticketemit-orch-metric.sh: has ticket + saved_est_tokensClassification pin: has("ticket") → companion; else if has("outcome") → run;
else skip. Never double-count. Sum saved_est_tokens only when numeric on
companion rows.
rollup.sh [--json] [--section all|local|council|outcomes|worktree]
Exit 0 success/partial; 64 usage. jq absent → degrade notice, exit 0.