メニュー
Full Meta-Harness status view — Pareto frontier, recent runs, regressions, eval health, installed plugins.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Evolve repo-local Claude Code harness assets through a 5-phase pipeline — harvest context, propose candidate, evaluate with evidence, audit regressions, report results.
Analyze the current project and generate initial eval tasks for harness optimization. Creates regression and capability eval tasks based on project structure.
Run the evaluation suite on the current harness or a specific candidate run. Reports deterministic check results and LLM-judge assessment.
Summarize the current Meta-Harness-style frontier of harness candidates, including quality, cost, latency, and safety notes.
Audit recent regressions in the harness search and identify likely confounds, brittle edits, and safer next-step ideas.
| name | dashboard |
| description | Full Meta-Harness status view — Pareto frontier, recent runs, regressions, eval health, installed plugins. |
| disable-model-invocation | true |
| allowed-tools | Read Grep Glob Bash(python3 *) Bash(mh-*) Bash(git *) |
mh-frontier --markdown 2>/dev/null || echo "No frontier data"
mh-regressions --markdown 2>/dev/null || echo "No regressions"
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/eval_runner.py --eval-dir ${CLAUDE_PLUGIN_ROOT}/eval-tasks --cwd . 2>&1 || echo "Eval runner unavailable"
python3 -c "
from scripts.meta_harness import detect_incomplete_runs
r = detect_incomplete_runs()
if r: print(f'Incomplete: {r[\"run_id\"]} phase={r[\"phase\"]} turn={r[\"turn\"]}')
else: print('No incomplete runs')
" 2>/dev/null || echo "Check unavailable"
python3 -c "
import json, pathlib
p = pathlib.Path.home() / '.claude' / 'plugins' / 'installed_plugins.json'
if p.exists():
d = json.loads(p.read_text())
for k in list(d.get('plugins', {}))[:10]:
name = k.split('@')[0]
print(f' {name}')
else:
print('No plugins')
" 2>/dev/null || echo "Plugin scan unavailable"
Present the dashboard using the Meta-Harness output style:
◉ DASHBOARD ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[Synthesize the data above into a clear status view]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Highlight: