Install any skill in seconds. Free to start, no credit card required.
Get Started Free →파이프라인 에이전트(specifier, planner, scheduler, builder, verifier, committer)가 참조하는 규칙/스키마 문서 모음.
.claude/skills/davepoon-sdd-pipeline-references/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -71% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -65% | 0% |
파이프라인 에이전트(specifier, planner, scheduler, builder, verifier, committer)가 참조하는 규칙/스키마 문서 모음.
직접 사용자 호출용이 아니라 에이전트 내부 참조용입니다.
| File | Description | |------|-------------| | references/agent-flow.md | 파이프라인 흐름 정의 | | references/file-content-schema.md | 파일 포맷 스키마 | | references/shared-prompt-sections.md | 공통 규칙 (TASK ID, PLAN.md 7 fields, WORK-LIST) | | references/context-policy.md | 슬라이딩 윈도우 규칙 | | references/xml-schema.md | XML 통신 포맷 | | references/work-activity-log.md | Activity Log 규칙 |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | pass→pass | 7,420 | 1,953 | -74% | 1 | 1 | 0% | 1,090 | 491 | -55% | 0 | 0 | — |
case-01 | fail→pass | 20,231 | 14,876 | -26% | 1 | 1 | 0% | 3,922 | 3,011 | -23% | 0 | 0 | — |
case-02 | fail→pass | 13,176 | 13,378 | +2% | 1 | 1 | 0% | 1,919 | 2,442 | +27% | 0 | 0 | — |
case-20 | pass→pass | 12,184 | 11,941 | -2% | 1 | 1 | 0% | 2,542 | 2,593 | +2% | 0 | 0 | — |
case-03 | pass→pass | 21,487 | 12,928 | -40% | 1 | 1 | 0% | 3,482 | 2,393 | -31% | 0 | 0 | — |
case-04 | fail→pass | 10,815 | 2,378 | -78% | 1 | 1 | 0% | 1,798 | 552 | -69% | 0 | 0 | — |
case-05 | fail→pass | 10,773 | 3,880 | -64% | 1 | 1 | 0% | 1,828 | 536 | -71% | 0 | 0 | — |
case-06 | fail→pass | 10,830 | 2,729 | -75% | 1 | 1 | 0% | 1,863 | 645 | -65% | 0 | 0 | — |
case-07 | fail→pass | 12,215 | 2,054 | -83% | 1 | 1 | 0% | 1,909 | 521 | -73% | 0 | 0 | — |
case-08 | fail→pass | 8,971 | 2,143 | -76% | 1 | 1 | 0% | 1,362 | 516 | -62% | 0 | 0 | — |
case-09 | fail→pass | 11,027 | 1,754 | -84% | 1 | 1 | 0% | 1,657 | 467 | -72% | 0 | 0 | — |
case-10 | fail→pass | 9,538 | 1,979 | -79% | 1 | 1 | 0% | 1,616 | 482 | -70% | 0 | 0 | — |
case-11 | fail→pass | 9,691 | 2,197 | -77% | 1 | 1 | 0% | 1,737 | 459 | -74% | 0 | 0 | — |
case-12 | fail→pass | 15,617 | 10,495 | -33% | 1 | 1 | 0% | 2,443 | 1,918 | -21% | 0 | 0 | — |
case-13 | fail→pass | 8,863 | 2,318 | -74% | 1 | 1 | 0% | 1,386 | 539 | -61% | 0 | 0 | — |
case-14 | pass→pass | 7,751 | 5,429 | -30% | 1 | 1 | 0% | 1,314 | 918 | -30% | 0 | 0 | — |
case-15 | pass→pass | 6,710 | 4,878 | -27% | 1 | 1 | 0% | 987 | 786 | -20% | 0 | 0 | — |
case-16 | pass→pass | 5,182 | 2,725 | -47% | 1 | 1 | 0% | 799 | 514 | -36% | 0 | 0 | — |
case-17 | pass→pass | 7,204 | 1,985 | -72% | 1 | 1 | 0% | 1,149 | 462 | -60% | 0 | 0 | — |
case-18 | fail→pass | 7,366 | 2,476 | -66% | 1 | 1 | 0% | 1,140 | 485 | -57% | 0 | 0 | — |
case-21 | pass→pass | 7,756 | 7,385 | -5% | 1 | 1 | 0% | 1,607 | 1,563 | -3% | 0 | 0 | — |
case-22 | pass→pass | 16,092 | 13,971 | -13% | 1 | 1 | 0% | 3,321 | 3,381 | +2% | 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 +59 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.