Install any skill in seconds. Free to start, no credit card required.
Get Started Free →- **Classification**: feature, architecture-change, hotfix (if touching API endpoints)
.claude/skills/bilal140202-kbwen-agentic-os/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -5% | 0% |
<!-- This is a SCAFFOLD skill. When /app-init runs in a downstream project, --> <!-- it customizes this file based on the project's ADR tech stack. --> <!-- If this file has NOT been customized, the AI should treat it as generic guidance. -->
> Customize after /app-init: Replace these generic conventions with your project's ADR decisions.
GET /users not GET /getUsers/users, /orders, /products/users/:id/orders/order-items| Action | Method | Success Status | Idempotent | |---|---|---|---| | List | GET | 200 | Yes | | Get one | GET | 200 | Yes | | Create | POST | 201 | No | | Full update | PUT | 200 | Yes | | Partial update | PATCH | 200 | No | | Delete | DELETE | 204 | Yes |
json{ "error": { "code": "VALIDATION_ERROR", "message": "Email is required", "details": { "field": "email", "constraint": "required" } } }
| HTTP Status | When | Error Code Pattern | |---|---|---| | 400 | Invalid input | VALIDATION_ERROR, INVALID_FORMAT | | 401 | Not authenticated | UNAUTHORIZED | | 403 | Not permitted | FORBIDDEN | | 404 | Resource not found | NOT_FOUND | | 409 | Conflict (duplicate) | CONFLICT, DUPLICATE | | 422 | Business rule violation | UNPROCESSABLE | | 429 | Rate limited | RATE_LIMITED | | 500 | Server error | INTERNAL_ERROR |
json{ "data": [...], "pagination": { "total": 100, "page": 1, "per_page": 20, "has_next": true } }
/api/v1/usersSunset header with date, keep old version for N monthsDuring /implement:
public annotation)During /review:
For phase-entry loading, read only:
When to ApplyChecklistLoad Conventions, Anti-Patterns, and References on full read or cache miss only.
POST /createUser → POST /users/user and /products (singular vs plural){ success: false }docs/adr/ADR-002-project-architecture.md § API Design.agent/rules/security_guardrails.md (A01: Access Control, A03: Injection).agentcortex/templates/spec-app-feature.md § API Contract| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 17,613 | 16,955 | -4% | 1 | 1 | 0% | 4,150 | 5,054 | +22% | 0 | 0 | — |
case-02 | fail→pass | 17,965 | 11,803 | -34% | 1 | 1 | 0% | 3,931 | 4,238 | +8% | 0 | 0 | — |
case-03 | pass→pass | 12,398 | 11,392 | -8% | 1 | 1 | 0% | 2,824 | 3,759 | +33% | 0 | 0 | — |
case-04 | pass→pass | 13,242 | 9,720 | -27% | 1 | 1 | 0% | 3,102 | 3,391 | +9% | 0 | 0 | — |
case-05 | pass→pass | 14,036 | 9,100 | -35% | 1 | 1 | 0% | 3,131 | 3,236 | +3% | 0 | 0 | — |
case-06 | pass→pass | 11,505 | 8,304 | -28% | 1 | 1 | 0% | 2,717 | 3,074 | +13% | 0 | 0 | — |
case-12 | fail→pass | 13,334 | 6,709 | -50% | 1 | 1 | 0% | 2,955 | 2,804 | -5% | 0 | 0 | — |
case-07 | pass→pass | 4,288 | 4,569 | +7% | 1 | 1 | 0% | 933 | 2,240 | +140% | 0 | 0 | — |
case-08 | pass→pass | 2,509 | 1,798 | -28% | 1 | 1 | 0% | 445 | 1,470 | +230% | 0 | 0 | — |
case-09 | fail→pass | 9,549 | 3,799 | -60% | 1 | 1 | 0% | 1,861 | 1,892 | +2% | 0 | 0 | — |
case-10 | fail→pass | 14,492 | 8,055 | -44% | 1 | 1 | 0% | 3,131 | 2,973 | -5% | 0 | 0 | — |
case-11 | fail→pass | 8,577 | 1,895 | -78% | 1 | 1 | 0% | 1,644 | 1,509 | -8% | 0 | 0 | — |
case-13 | pass→pass | 13,766 | 9,889 | -28% | 1 | 1 | 0% | 2,612 | 3,100 | +19% | 0 | 0 | — |
case-14 | pass→pass | 12,202 | 2,714 | -78% | 1 | 1 | 0% | 2,128 | 1,648 | -23% | 0 | 0 | — |
case-15 | pass→pass | 9,227 | 5,208 | -44% | 1 | 1 | 0% | 1,944 | 2,166 | +11% | 0 | 0 | — |
case-16 | fail→pass | 8,367 | 1,576 | -81% | 1 | 1 | 0% | 1,701 | 1,444 | -15% | 0 | 0 | — |
case-17 | pass→pass | 7,568 | 4,584 | -39% | 1 | 1 | 0% | 1,632 | 2,095 | +28% | 0 | 0 | — |
case-18 | pass→pass | 12,732 | 6,366 | -50% | 1 | 1 | 0% | 2,354 | 2,315 | -2% | 0 | 0 | — |
case-19 | fail→fail | 8,675 | 3,450 | -60% | 1 | 1 | 0% | 1,601 | 1,795 | +12% | 0 | 0 | — |
case-20 | fail→pass | 8,180 | 2,057 | -75% | 1 | 1 | 0% | 1,458 | 1,539 | +6% | 0 | 0 | — |
case-21 | pass→pass | 10,543 | 4,302 | -59% | 1 | 1 | 0% | 1,967 | 1,861 | -5% | 0 | 0 | — |
case-22 | pass→pass | 7,958 | 2,480 | -69% | 1 | 1 | 0% | 1,539 | 1,627 | +6% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.