prd-executor — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited prd-executor (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill guides a Dev Agent through executing a prd_decomposer breakdown located at docs/[Feature_Name]_breakdown/. It enforces three hard gates:
docs/[Feature_Name]_breakdown/Read docs/[Feature_Name]_breakdown/outline.md. Extract:
Announce to the user: which REQs are pending, which (if any) are already in progress or done.
Find the first Phase in outline.md whose status is ⬜ (not started) and whose REQ dependencies (if any cross-Phase) are all ✅ or ⚠️.
If no such Phase exists:
outline.md to 💬 (Discussing)a. Open the REQ file. Check whether 🔍 详细需求确认 still contains ⚠️ [待填充]. b. If YES (placeholder present):
outline.md to 💬superpowers:brainstorming with the REQ content as context⚠️ [待填充] placeholder block with the brainstorming conclusionsc. If NO (placeholder already replaced): skip brainstorming for this REQ.
outline.md to 🔍 (Pending Confirmation) Phase X Discussion Summary:
- REQ-X.1 [Title]: [2-3 key conclusions]
- REQ-X.2 [Title]: [2-3 key conclusions]
- REQ-X.3 [Title]: [2-3 key conclusions]outline.md to 🚧 (Implementing)a. Update REQ status in REQ file and outline.md to 🚧 (In Progress) b. Invoke superpowers:test-driven-development with the REQ's Acceptance Criteria as test targets c. When Acceptance Criteria pass internally: keep REQ status as 🚧 — it will be updated to ✅ in Stage G after the user passes UAT d. Commit immediately after each REQ:
git add -p # stage only implementation files, NOT debug/mock artifacts
git commit -m "feat(REQ-X.X): [short description of what was built]"outline.md to 🧪 (Pending UAT) Phase X UAT Checklist:
REQ-X.1 [Title]
Please verify:
- [ ] [AC item 1]
- [ ] [AC item 2]
REQ-X.2 [Title]
Please verify:
- [ ] [AC item 1]
- [ ] [AC item 2]Wait for the user to report results. While waiting, keep Phase status as 🧪 (Pending UAT) and do not modify any code. Expected format (user can use any natural language):
For each REQ the user reports as failing:
REQ-X.2 [Title] — Fixed. Please re-verify:
- [ ] [AC item 1]
- [ ] [AC item 2]When all REQs in the Phase have passed UAT:
outline.md最后更新 field in each REQ file to today's dateoutline.md to ✅ (Complete) git push| Status | Emoji | Meaning |
|---|---|---|
| Not Started | ⬜ | Default at decomposition time |
| Discussing | 💬 | Brainstorming in progress |
| Changed | 🔄 | Requirements modified after confirmation (update 备注 with what changed and why) |
| In Progress | 🚧 | Implementation underway |
| Done | ✅ | All Acceptance Criteria verified |
| Partial | ⚠️ | Main feature done, some AC items shelved (reason in 备注) |
| Shelved | 🚫 | Blocked, see 备注 for reason |
| Status | Emoji | Meaning |
|---|---|---|
| Not Started | ⬜ | Default |
| Discussing | 💬 | Brainstorming REQs in progress |
| Pending Confirmation | 🔍 | All REQs discussed, waiting for user sign-off |
| Implementing | 🚧 | AI implementing all REQs autonomously |
| Pending UAT | 🧪 | Waiting for user verification |
| Complete | ✅ | All REQs passed UAT, pushed |
If requirements change mid-execution on a REQ that is 🚧 In Progress:
🎯 Requirement and ✅ Acceptance Criteria sections of the REQ fileoutline.md; both must stay in sync~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.