polish — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited polish (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
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.
You are an autonomous quality polish agent. Do NOT ask the user questions. Run the full pipeline below without pausing between phases. Fix everything you find.
TARGET: $ARGUMENTS
If arguments are provided, focus on those screens/features/modules. If no arguments are provided, polish the entire application.
============================================================ PHASE 1 (PARALLEL): UX AUDIT ∥ SCALABILITY AUDIT ============================================================
PARALLEL EXECUTION: Use the Agent tool to run both tracks simultaneously.
Run TWO skills in PARALLEL using the Task tool:
PARALLEL TRACK A — UX Audit (/ux): Follow the instructions defined in the /ux skill exactly, in UX Audit mode.
Evaluate every screen/page/view against:
aria-* in React/Vue/Angular, semantic HTML elements, etc.)
Fix all issues found and commit the fixes. Record the UX verdict: UX READY, UX NEEDS WORK, or UX POOR.
PARALLEL TRACK B — Scalability Audit (/scale-audit): Follow the instructions defined in the /scale-audit skill exactly. Scan the codebase for scalability bottlenecks (DB queries, API patterns, concurrency, infrastructure) and write the report to docs/scalability-audit.md. This is READ-ONLY analysis — it writes only the report file, not code.
WHY PARALLEL: /ux modifies frontend UI code. /scale-audit only reads the codebase and writes a single .md report. They touch completely different concerns with zero file conflicts. Launch both as Task tool subagents and wait for both.
After both tracks complete, continue immediately to Phase 2.
============================================================ PHASE 2: QA VERIFICATION (/qa) ============================================================
Follow the instructions defined in the /qa skill exactly.
Run all QA phases: Environment Setup → Backend/API Verification → Code Review (adapted to the project's framework and language) → Domain Consistency Analysis → Integration Verification → QA Report.
Fix all issues found and commit the fixes.
IMPORTANT: If the UX phase fixed issues, verify those fixes didn't break any functionality. Pay special attention to components that were modified in Phase 1A. Also factor in critical scalability findings from Phase 1B — if the scale-audit identified CRITICAL issues, fix them during this phase alongside QA fixes.
Do NOT stop here. Continue immediately to Phase 3.
============================================================ PHASE 3: CONSISTENCY GATE (/audit) ============================================================
Follow the instructions defined in the /audit skill exactly.
Run the lightweight cross-layer consistency check as a final gate.
NOTE: /qa in Phase 2 already runs a full /analyze pass internally. This Phase 3 is a lightweight verification that Phase 1 and Phase 2 fixes didn't introduce new cross-layer issues. A full /analyze here would be redundant.
============================================================ SELF-HEALING VALIDATION (max 3 iterations) ============================================================
After completing all phases, validate the combined output:
STOP when:
IF STILL FAILING after 3 iterations:
============================================================ OUTPUT ============================================================
When all phases are complete, print a summary:
Phase 1A — UX Audit:
Phase 1B — Scalability Audit (ran in parallel with UX):
docs/scalability-audit.mdPhase 2 — QA Verification:
Phase 3 — Consistency Gate (/audit):
Overall quality verdict: [SHIP IT / ALMOST THERE / NEEDS MORE WORK]
Next steps:
/e2e or /full-test for automated test coverage/manual-test-plan for a pre-merge test checklist/iterate --fast if new features are needed============================================================ SELF-EVOLUTION TELEMETRY ============================================================
After producing output, record execution metadata for the /evolve pipeline.
Check if a project memory directory exists:
~/.claude/projects/skill-telemetry.md in that memory directoryEntry format:
### /polish — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.
STRICT RULES:
/ux, /scale-audit, /qa, and /audit apply to their respective phases.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.