メニュー
Use when reviewing implementation work that ran through a /tmp Pi task harness. Checks contract drift, unsafe shortcuts, validation evidence, and status honesty.
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
Use when running an existing /tmp Pi task harness. Executes inside the active contract, records status, respects enforcement, and loops through validation.
Use when creating executable /tmp Pi task harnesses for ambiguous or high-risk coding work. Builds contract, prompts, validation, logs, and enforcement rules.
Use when developing or reviewing packages/pi-observability dashboard/server changes locally, especially rebuilding dist, restarting the 43190 server, validating UI behavior, or keeping the local dashboard in sync with source edits.
Use when evaluating proposed Pi customisation work against built-in Pi primitives before implementation, especially extensions, packages, slash commands, and workflow ideas.
Use when creating or modifying Pi extensions, built-in extension wiring, prompt shims, package scaffolds, package READMEs, or packages/pi-* implementation in this monorepo.
Use when verifying that published Pi packages install and load for normal users, including pi.dev package compatibility checks, release validation, Daytona sandbox tests, and package installability audits.
| name | review-harness |
| description | Use when reviewing implementation work that ran through a /tmp Pi task harness. Checks contract drift, unsafe shortcuts, validation evidence, and status honesty. |
| compatibility | Requires my-pi or Pi with the pi-harness extension tools enabled. |
Review the completed or failed work against the harness contract.
harness.json, TASK.md,
status.json, OUTCOME.md or outcome.json, and
logs/events.jsonl if present.<harness_dir>/review.sh.allowed_paths, forbidden_paths,
test-change policy, and escalation rules.status.json, logs, and command
output.allow_test_changes is false?## Verdict
Pass | Needs executor fix | Escalate
## Findings
- P0/P1/P2: <issue> — `<path>` — <required action>
## Validation Evidence
- `<command>` — pass/fail/not run — <evidence>
## Required Next Step
- <one action>