Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Keeps implementation and specs in sync. Use when working on a feature that has a spec in .agents/specs/, when the user says /spec, or when starting implementation of a documented feature. Also use when the user asks to verify implementation against a spec or update a spec after changes.
.claude/skills/n8n-io-n8n-spec-driven-development/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -23% | 0% |
Specs live in .agents/specs/. They are the source of truth for architectural decisions, API contracts, and implementation scope. Implementation and specs must stay in sync — neither leads exclusively.
Read spec → Implement → Verify alignment → Update spec or code → Repeat.agents/specs/ for files matching the feature:bashls .agents/specs/
open questions before writing code.
architectural change), ask the user whether to create one first.
change, constraint discovered), update the spec immediately in the same session. Don't leave spec and code out of sync.
- [ ] → - [x]) as items are completed.replaced, with a brief reason: markdown
Run a spec verification pass:
@n8n/api-types?implementation doesn't deliver yet (acceptable for MVP, but should be noted).
.agents/specs/.## Implementation TODO with checkboxes to track progress.but don't enforce a rigid naming scheme.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,279 | 7,838 | +48% | 1 | 1 | 0% | 252 | 1,030 | +309% | 0 | 0 | — |
case-02 | fail→fail | 11,865 | 5,834 | -51% | 1 | 1 | 0% | 202 | 979 | +385% | 0 | 0 | — |
case-03 | fail→fail | 7,697 | 5,877 | -24% | 1 | 1 | 0% | 193 | 891 | +362% | 0 | 0 | — |
case-04 | fail→fail | 8,096 | 8,274 | +2% | 1 | 1 | 0% | 679 | 1,211 | +78% | 0 | 0 | — |
case-05 | pass→fail | 12,628 | 7,603 | -40% | 1 | 1 | 0% | 1,836 | 924 | -50% | 0 | 0 | — |
case-06 | pass→fail | 29,413 | 6,755 | -77% | 1 | 1 | 0% | 6,352 | 835 | -87% | 0 | 0 | — |
case-07 | pass→pass | 4,761 | 5,425 | +14% | 1 | 1 | 0% | 689 | 1,413 | +105% | 0 | 0 | — |
case-08 | fail→pass | 9,013 | 4,427 | -51% | 1 | 1 | 0% | 1,283 | 1,136 | -11% | 0 | 0 | — |
case-09 | pass→pass | 11,417 | 5,021 | -56% | 1 | 1 | 0% | 1,612 | 1,413 | -12% | 0 | 0 | — |
case-10 | fail→pass | 13,847 | 2,572 | -81% | 1 | 1 | 0% | 2,254 | 986 | -56% | 0 | 0 | — |
case-11 | pass→pass | 10,452 | 3,487 | -67% | 1 | 1 | 0% | 1,244 | 1,154 | -7% | 0 | 0 | — |
case-12 | fail→pass | 8,740 | 2,468 | -72% | 1 | 1 | 0% | 1,363 | 901 | -34% | 0 | 0 | — |
case-13 | fail→pass | 10,695 | 2,952 | -72% | 1 | 1 | 0% | 1,515 | 990 | -35% | 0 | 0 | — |
case-14 | pass→pass | 8,128 | 3,690 | -55% | 1 | 1 | 0% | 1,212 | 1,046 | -14% | 0 | 0 | — |
case-15 | pass→pass | 10,211 | 5,292 | -48% | 1 | 1 | 0% | 1,563 | 1,445 | -8% | 0 | 0 | — |
case-16 | fail→pass | 10,492 | 3,081 | -71% | 1 | 1 | 0% | 1,392 | 1,075 | -23% | 0 | 0 | — |
case-17 | pass→pass | 10,459 | 7,231 | -31% | 1 | 1 | 0% | 1,486 | 1,731 | +16% | 0 | 0 | — |
case-18 | pass→pass | 15,203 | 8,946 | -41% | 1 | 1 | 0% | 2,214 | 1,810 | -18% | 0 | 0 | — |
case-19 | pass→pass | 2,639 | 3,443 | +30% | 1 | 1 | 0% | 298 | 1,106 | +271% | 0 | 0 | — |
case-20 | pass→fail | 10,776 | 5,648 | -48% | 1 | 1 | 0% | 1,502 | 1,461 | -3% | 0 | 0 | — |
case-21 | pass→pass | 13,449 | 4,322 | -68% | 1 | 1 | 0% | 1,823 | 1,286 | -29% | 0 | 0 | — |
case-22 | pass→pass | 15,092 | 4,332 | -71% | 1 | 1 | 0% | 2,318 | 1,280 | -45% | 0 | 0 | — |
case-23 | pass→pass | 7,487 | 3,161 | -58% | 1 | 1 | 0% | 1,114 | 1,021 | -8% | 0 | 0 | — |
case-24 | fail→pass | 8,138 | 2,195 | -73% | 1 | 1 | 0% | 1,066 | 843 | -21% | 0 | 0 | — |
case-25 | fail→fail | 12,495 | 4,137 | -67% | 1 | 1 | 0% | 1,672 | 1,219 | -27% | 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. 25 cases were attempted, and 19 counted toward the lift figure. The other 6 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +12 percentage points is the difference between those two pass rates over the 19 comparable cases. 4 cases got worse with the skill loaded, and they are included in that figure.
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.