Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate comprehensive C4 architecture documentation for an existing repository/codebase using a bottom-up analysis approach.
.claude/skills/dokhacgiakhoa-c4-architecture-c4-architecture/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 33% | 0% |
Generate comprehensive C4 architecture documentation for an existing repository/codebase using a bottom-up analysis approach.
Extended thinking: This workflow implements a complete C4 architecture documentation process following the C4 model (Context, Container, Component, Code). It uses a bottom-up approach, starting from the deepest code directories and working upward, ensuring every code element is documented before synthesizing into higher-level abstractions. The workflow coordinates four specialized C4 agents (Code, Component, Container, Context) to create a complete architectural documentation set that serves both technical and non-technical stakeholders.]
resources/implementation-playbook.md.This workflow creates comprehensive C4 architecture documentation following the official C4 model by:
Note: According to the C4 model, you don't need to use all 4 levels of diagram - the system context and container diagrams are sufficient for most software development teams. This workflow generates all levels for completeness, but teams can choose which levels to use.
All documentation is written to a new C4-Documentation/ directory in the repository root.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 2,493 | 2,099 | -16% | 1 | 1 | 0% | 346 | 1,077 | +211% | 0 | 0 | — |
case-01 | fail→fail | 17,956 | 5,110 | -72% | 1 | 1 | 0% | 3,353 | 1,330 | -60% | 0 | 0 | — |
case-03 | fail→pass | 19,297 | 5,311 | -72% | 1 | 1 | 0% | 3,619 | 1,233 | -66% | 0 | 0 | — |
case-04 | pass→pass | 10,828 | 9,650 | -11% | 1 | 1 | 0% | 1,870 | 2,496 | +33% | 0 | 0 | — |
case-05 | fail→pass | 11,842 | 10,012 | -15% | 1 | 1 | 0% | 1,983 | 2,502 | +26% | 0 | 0 | — |
case-06 | fail→fail | 11,041 | 7,588 | -31% | 1 | 1 | 0% | 2,016 | 2,160 | +7% | 0 | 0 | — |
case-07 | pass→pass | 13,436 | 7,333 | -45% | 1 | 1 | 0% | 1,965 | 2,020 | +3% | 0 | 0 | — |
case-08 | pass→pass | 11,274 | 8,703 | -23% | 1 | 1 | 0% | 1,873 | 2,220 | +19% | 0 | 0 | — |
case-09 | fail→pass | 9,209 | 5,005 | -46% | 1 | 1 | 0% | 1,486 | 1,673 | +13% | 0 | 0 | — |
case-10 | pass→pass | 11,779 | 14,063 | +19% | 1 | 1 | 0% | 2,144 | 2,906 | +36% | 0 | 0 | — |
case-11 | pass→pass | 11,972 | 8,611 | -28% | 1 | 1 | 0% | 1,768 | 2,217 | +25% | 0 | 0 | — |
case-12 | pass→pass | 15,200 | 8,501 | -44% | 1 | 1 | 0% | 2,169 | 2,039 | -6% | 0 | 0 | — |
case-13 | pass→pass | 14,285 | 12,130 | -15% | 1 | 1 | 0% | 2,351 | 2,693 | +15% | 0 | 0 | — |
case-14 | fail→pass | 14,442 | 7,431 | -49% | 1 | 1 | 0% | 1,947 | 1,857 | -5% | 0 | 0 | — |
case-15 | pass→pass | 13,195 | 7,606 | -42% | 1 | 1 | 0% | 1,851 | 2,024 | +9% | 0 | 0 | — |
case-16 | pass→pass | 6,232 | 3,580 | -43% | 1 | 1 | 0% | 907 | 1,359 | +50% | 0 | 0 | — |
case-17 | pass→pass | 10,458 | 9,485 | -9% | 1 | 1 | 0% | 1,779 | 2,321 | +30% | 0 | 0 | — |
case-18 | fail→fail | 6,537 | 7,190 | +10% | 1 | 1 | 0% | 1,044 | 1,924 | +84% | 0 | 0 | — |
case-19 | pass→pass | 6,594 | 3,466 | -47% | 1 | 1 | 0% | 1,093 | 1,246 | +14% | 0 | 0 | — |
case-20 | fail→fail | 17,103 | 13,470 | -21% | 1 | 1 | 0% | 3,480 | 3,236 | -7% | 0 | 0 | — |
case-21 | fail→fail | 20,340 | 18,269 | -10% | 1 | 1 | 0% | 3,005 | 3,519 | +17% | 0 | 0 | — |
case-22 | fail→fail | 24,884 | 19,125 | -23% | 1 | 1 | 0% | 3,318 | 3,471 | +5% | 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 +18 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.