Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generates a Mermaid dependency graph showing import relationships between modules. Use when analyzing coupling, finding circular deps, or planning refactors.
.claude/skills/athola-dependency-graph/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 2% | 0% |
| case-07 | ✓→✗ | ▼ Worse | -26% | 0% |
| case-08 | ✓→✗ | ▼ Worse | -44% | 0% |
Generate a Mermaid flowchart showing import and dependency relationships between modules, packages, or plugins.
cartograph:call-chain)cartograph:class-diagram)Dispatch the codebase explorer agent:
Agent(cartograph:codebase-explorer)
Prompt: Explore [scope] and return a structural model.
Focus on import statements and cross-module dependencies
for a dependency graph. Track both internal and external
imports.Transform the structural model into a Mermaid flowchart with directed edges representing dependencies.
Rules for dependency graphs:
flowchart LR (left-right) for dependency direction"A depends on B")
-.->) for external/optional deps==>) for critical path dependenciesExample output:
mermaidflowchart LR subgraph sanctum[Sanctum] commit[commit] pr_prep[pr_prep] workspace[workspace] end subgraph leyline[Leyline] git[git_platform] errors[error_patterns] end subgraph external[External] subprocess[subprocess] json[json] end commit --> git commit --> json pr_prep --> workspace pr_prep --> git workspace --> errors workspace -.-> subprocess
Call the Mermaid Chart MCP to render:
mcp__claude_ai_Mermaid_Chart__validate_and_render_mermaid_diagram
prompt: "Dependency graph of [scope]"
mermaidCode: [generated syntax]
diagramType: "flowchart"
clientName: "claude-code"If rendering fails, fix syntax and retry (max 2 retries).
Show the rendered diagram with analysis notes:
flowchart LR syntax generated with directed edgesfrom dependent to dependency (A --> B means "A depends on B")
mcp__claude_ai_Mermaid_Chart__validate_and_render_mermaid_diagramcalled and returns without error
and highest fan-out modules
styling and listed explicitly in the analysis notes
-.->) distinct frominternal dependency arrows
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 19,634 | 6,737 | -66% | 1 | 1 | 0% | 3,464 | 1,114 | -68% | 0 | 0 | — |
case-02 | fail→fail | 3,878 | 31,646 | +716% | 1 | 1 | 0% | 197 | 7,289 | +3600% | 0 | 0 | — |
case-03 | fail→fail | 5,027 | 7,268 | +45% | 1 | 1 | 0% | 276 | 1,458 | +428% | 0 | 0 | — |
case-04 | pass→fail | 13,017 | 9,734 | -25% | 1 | 1 | 0% | 2,186 | 2,238 | +2% | 0 | 0 | — |
case-05 | pass→pass | 17,382 | 23,263 | +34% | 1 | 1 | 0% | 3,057 | 5,325 | +74% | 0 | 0 | — |
case-06 | pass→pass | 12,328 | 7,101 | -42% | 1 | 1 | 0% | 1,848 | 1,810 | -2% | 0 | 0 | — |
case-07 | pass→fail | 9,920 | 6,907 | -30% | 1 | 1 | 0% | 1,640 | 1,206 | -26% | 0 | 0 | — |
case-08 | pass→fail | 12,516 | 6,606 | -47% | 1 | 1 | 0% | 2,150 | 1,212 | -44% | 0 | 0 | — |
case-09 | pass→fail | 13,777 | 15,780 | +15% | 1 | 1 | 0% | 2,327 | 1,281 | -45% | 0 | 0 | — |
case-10 | pass→fail | 12,506 | 6,412 | -49% | 1 | 1 | 0% | 2,534 | 1,148 | -55% | 0 | 0 | — |
case-11 | pass→fail | 9,632 | 32,323 | +236% | 1 | 1 | 0% | 1,609 | 1,013 | -37% | 0 | 0 | — |
case-12 | fail→pass | 11,223 | 26,201 | +133% | 1 | 1 | 0% | 2,045 | 3,137 | +53% | 0 | 0 | — |
case-13 | fail→fail | 5,871 | 6,697 | +14% | 1 | 1 | 0% | 233 | 1,064 | +357% | 0 | 0 | — |
case-14 | pass→fail | 15,001 | 58,222 | +288% | 1 | 1 | 0% | 2,659 | 7,261 | +173% | 0 | 0 | — |
case-15 | pass→fail | 12,875 | 9,232 | -28% | 1 | 1 | 0% | 2,155 | 1,133 | -47% | 0 | 0 | — |
case-16 | pass→pass | 11,983 | 2,105 | -82% | 1 | 1 | 0% | 1,826 | 1,098 | -40% | 0 | 0 | — |
case-17 | pass→pass | 9,518 | 1,609 | -83% | 1 | 1 | 0% | 1,461 | 982 | -33% | 0 | 0 | — |
case-18 | pass→pass | 4,890 | 2,379 | -51% | 1 | 1 | 0% | 872 | 1,086 | +25% | 0 | 0 | — |
case-19 | pass→fail | 20,844 | 61,858 | +197% | 1 | 1 | 0% | 4,336 | 7,094 | +64% | 0 | 0 | — |
case-20 | pass→fail | 15,758 | 6,525 | -59% | 1 | 1 | 0% | 3,283 | 1,049 | -68% | 0 | 0 | — |
case-21 | pass→fail | 13,829 | 7,915 | -43% | 1 | 1 | 0% | 2,337 | 1,329 | -43% | 0 | 0 | — |
case-22 | fail→pass | 7,043 | 1,536 | -78% | 1 | 1 | 0% | 1,024 | 961 | -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, and 8 counted toward the lift figure. The other 14 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 -41 percentage points is the difference between those two pass rates over the 8 comparable cases. 11 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.