菜单
Architecture drift detection, silent failure scanning, hallucination prevention, and feature completeness auditing. Activates at phase boundaries, before merges, or on explicit audit requests.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/CleanExpo/DR-NRPG --skill system-supervisor命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Validate Australian phone numbers, addresses, postcodes, ABN/ACN numbers, and state/territory references. Use when reviewing any content with Australian business data.
Domain knowledge for Australian disaster recovery and restoration services. Use when working on content, service descriptions, or any feature related to the DR/NRPG business model.
Verify factual accuracy of all claims, data, and information across the platform. Use when reviewing content, service descriptions, or business information for correctness.
正在显示 SKILL.md
基于 SOC 职业分类
| name | system-supervisor |
| description | Architecture drift detection, silent failure scanning, hallucination prevention, and feature completeness auditing. Activates at phase boundaries, before merges, or on explicit audit requests. |
| license | MIT |
| metadata | {"author":"NodeJS-Starter-V1","version":"1.0.0","locale":"en-AU"} |
Detects specification drift, silent failures, hallucinated assertions, and incomplete features. Provides a post-execution audit layer that complements the Execution Guardian's pre-execution gates.
Scans the codebase for architectural integrity issues that accumulate over time. Compares implementation against specifications, detects dead code and silent failures, validates technical assertions, and measures feature completeness across all layers. Distinct from the Execution Guardian (pre-execution safety) and the Council of Logic (code quality).
docs/phases/, docs/features/, and spec files for declared features, endpoints, models, and components| Category | Meaning | Severity |
|---|---|---|
| MISSING | Spec declares it, codebase does not implement it | HIGH |
| ORPHANED | Codebase implements it, no spec references it | MEDIUM |
| DIVERGED | Both exist but implementation differs from spec | HIGH |
| ALIGNED | Spec and implementation match | OK |
## Architecture Drift Report — {date}
| Item | Spec Location | Code Location | Status | Severity |
|------|--------------|---------------|--------|----------|
| User authentication | docs/features/auth/spec.md | apps/backend/src/auth/ | ALIGNED | OK |
| Contractor API | docs/phases/phase-2-spec.md | apps/backend/src/api/contractors.py | DIVERGED | HIGH |
| Analytics dashboard | docs/features/analytics/spec.md | — | MISSING | HIGH |
| Legacy helper | — | apps/backend/src/utils/old_helper.py | ORPHANED | MEDIUM |
### Summary
- ALIGNED: {n}
- MISSING: {n} (HIGH priority)
- ORPHANED: {n} (review for removal)
- DIVERGED: {n} (reconcile spec or code)
See references/drift-rules.md for:
Silent failures are code constructs that fail without alerting. They accumulate technical debt invisibly.
| Silent Failure Type | Detection Pattern | Severity |
|---|---|---|
| Dead imports | import X where X is never used in the file | LOW |
| Empty catch blocks | except: or catch {} with no logging or re-raise | HIGH |
| Unused API endpoints | Route defined but no frontend calls reference it | MEDIUM |
| Orphaned DB columns | Column in model but never read/written by any query | MEDIUM |
| Unchecked return values | Async function called without await or return value discarded | HIGH |
| Unreachable code paths | Code after unconditional return, raise, or break | LOW |
| Stale environment variables | .env.example references variable not used in code | LOW |
| Unhandled promise rejections | .then() without .catch() or missing error boundary | HIGH |
| Type assertion bypasses | as any, # type: ignore without justification comment | MEDIUM |
| Commented-out code blocks | More than 5 consecutive commented lines of code | LOW |
## Silent Failure Scan — {date}
| # | Type | Location | Line | Severity | Recommendation |
|---|------|----------|------|----------|---------------|
| 1 | Empty catch | apps/backend/src/api/main.py | 47 | HIGH | Add logging or re-raise |
| 2 | Dead import | apps/web/lib/api/client.ts | 3 | LOW | Remove unused import |
| 3 | Unchecked await | apps/backend/src/agents/base.py | 92 | HIGH | Add await or handle return |
### Summary
- HIGH: {n} (fix immediately)
- MEDIUM: {n} (fix before merge)
- LOW: {n} (fix when convenient)
AI coding agents can make assertions about code behaviour that are not verified. These assertions become "technical hallucinations" — statements treated as fact that may be incorrect.
When the agent makes a technical assertion (e.g., "this function handles errors correctly", "this endpoint returns 404"), classify it:
| Classification | Meaning | Action |
|---|---|---|
| CONFIRMED | Verified by test, code inspection, or documentation | No action needed |
| INFERRED | Logically follows from confirmed facts but not directly verified | Acceptable with note |
| ASSUMED | Plausible but not verified | Trigger verification pass |
| FABRICATED | No evidence supports the assertion | Block and correct immediately |
When an assertion is classified as ASSUMED:
## Assertion Verification — {date}
| # | Assertion | Classification | Evidence | Action |
|---|-----------|---------------|----------|--------|
| 1 | "Auth middleware validates JWT expiry" | CONFIRMED | test_auth.py:test_expired_token passes | None |
| 2 | "Rate limiter returns 429 after 100 req/min" | ASSUMED | No rate limiter test exists | Write test |
| 3 | "Contractor API returns paginated results" | FABRICATED | API returns full list, no pagination | Correct claim |
This hallucination check applies to technical code assertions only. For content-level claims (marketing copy, user-facing text), defer to content review processes.
For each declared feature, verify implementation across all required layers:
| Layer | What to Check | Path Pattern |
|---|---|---|
| UI | React component exists and renders | apps/web/app/**/*.tsx, apps/web/components/**/*.tsx |
| API Route | Backend endpoint defined | apps/backend/src/api/**/*.py |
| Data Model | SQLAlchemy model or Pydantic schema | apps/backend/src/db/**/*.py, apps/backend/src/models/**/*.py |
| Validation | Input validation (Zod frontend, Pydantic backend) | Inline in route handlers and components |
| Tests | At least one test per layer | apps/backend/tests/**/*.py, apps/web/**/*.test.{ts,tsx} |
| Docs | Feature documented | docs/features/**/*.md, docs/reference/**/*.md |
| Error Handling | Errors follow error-taxonomy patterns | Inline in route handlers |
## Feature Completeness — {date}
| Feature | UI | API | Model | Validation | Tests | Docs | Errors | Score |
|---------|----|----|-------|-----------|-------|------|--------|-------|
| Auth (login/logout) | Y | Y | Y | Y | Y | Y | Y | 100% |
| Contractor profiles | Y | Y | Y | P | N | N | P | 57% |
| Document management | Y | Y | Y | Y | P | Y | Y | 86% |
| Analytics dashboard | N | P | N | N | N | N | N | 14% |
**Legend**: Y = Complete, P = Partial, N = Missing
### Thresholds
- 100%: Release-ready
- 80-99%: Acceptable for merge (document gaps)
- 50-79%: Requires completion plan before merge
- <50%: Not ready — must complete core layers first
See references/completeness-matrix.md for detailed layer-by-layer checklists and path patterns.
Portable, project-scoped signals only. No investor readiness, competitive analysis, or business metrics — those are non-portable and outside this skill's scope.
Track across audit runs:
## Technical Debt Signals
| Metric | Previous | Current | Trend |
|--------|----------|---------|-------|
| Silent failures (HIGH) | 12 | 8 | Improving |
| Architecture drift items | 5 | 7 | Worsening |
| Feature completeness avg | 72% | 78% | Improving |
| Type assertion bypasses | 3 | 1 | Improving |
| Empty catch blocks | 6 | 4 | Improving |
## Dependency Health
| Package | Current | Latest | Status |
|---------|---------|--------|--------|
| next | 15.x.x | 15.x.x | Current |
| fastapi | 0.x.x | 0.x.x | Current |
| sqlalchemy | 2.x.x | 2.x.x | Current |
| {package} | {ver} | {ver} | Outdated (n versions behind) |
Overall completion: {n}% ({completed}/{total} features at 80%+ threshold)
Run all four scans in sequence:
Run abbreviated scans:
Run individual scans based on user request:
| Pattern | Problem | Correct Approach |
|---|---|---|
| Running full audit during active coding | Interrupts flow, wastes tokens | Audit at phase boundaries or on request |
| Flagging all ORPHANED code as errors | Some utilities are intentionally decoupled | Check if referenced in tests or docs before flagging |
| Treating INFERRED assertions as FABRICATED | Over-verification wastes effort | INFERRED is acceptable when logic chain is clear |
| Auditing generated/config files for drift | False positives from auto-generated content | Apply ignore patterns from drift-rules.md |
| Running Strategic Intelligence every scan | Token-heavy, low signal frequency | Run at milestone boundaries only |
docs/phases/ and docs/features/[AGENT_ACTIVATED]: System Supervisor
[MODE]: {BUILD | SCALE}
[SCAN_TYPE]: {full | quick | on-demand}
[STATUS]: {scanning | complete}
{audit reports}
[NEXT_ACTION]: {proceed with merge | fix {n} issues | schedule follow-up}