メニュー
Use when entering VERIFYING or before GATING and fresh evidence is required. Decides which evidence types the task needs, then runs collect_evidence.py; MUST NOT judge pass/fail itself.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
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。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/yezhwi/superpowers-engineering-harness --skill collect-evidenceコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Use when starting or resuming a task in a repo with .harness/ state, when unsure which harness phase comes next, or when asked to run the Engineering Harness loop end-to-end. Orchestrates the fixed state machine; MUST NOT implement business code itself.
Review Q2/Q3 changed business paths for production diagnosability and persist structured review input.
Turn a User Requirement into a Task Contract: Acceptance Criteria, Requirements, Invariants, Risks, Verification Plan. Produces .harness/requirements.yaml and .harness/invariants.yaml and advances task state CREATED -> SPECIFYING -> PLANNED. This skill MUST NOT implement business code.
| name | collect-evidence |
| description | Use when entering VERIFYING or before GATING and fresh evidence is required. Decides which evidence types the task needs, then runs collect_evidence.py; MUST NOT judge pass/fail itself. |
You decide WHAT evidence the task requires. collect_evidence.py does the
measuring.
From the task contract's Verification Plan + what the diff touched:
task modifies backend API
→ required: build, unit_test, integration_test
Map each required verification in .harness/requirements.yaml /
current-task.yaml to an evidence type. If a required verification lacks a
matching type, that is a spec problem — surface it, do not silently skip.
For each required type:
harness evidence --type unit_test --command "pytest"
(Inside the harness repo itself, the equivalent raw call is
python scripts/collect_evidence.py --type ... --command ...; elsewhere
always use the CLI.)
Generates .harness/evidence/<type>.json with command, exit_code, timestamp,
commit, stdout_tail, stderr_tail. Failing commands still produce evidence —
that is by design; never re-run until green to hide a failure.
collect_evidence.py output counts..harness/evidence/.