Skip to content

mgiovani/ai-arsenal / skills/docs-diagram

docs-diagram

Verified

Generate a Mermaid diagram (ER, architecture, deployment, or security) by reading the actual codebase and writing it to docs/. Use for requests like "draw an ER diagram of these tables", "show me the architecture", "diagram the deployment setup", "visualize the security data flow", or "generate a system diagram". Not for scaffolding a new docs directory structure (use docs-init), syncing existing prose docs with code changes (use docs-update), or recording a design decision's rationale (use docs-adr), this skill only produces diagrams generated from code.

evals/evals.json

View on GitHub
{
  "skill": "docs-diagram",
  "evals": [
    {
      "id": "happy-path-scoped-er",
      "prompt": "In this repo there's a SQLAlchemy models file at app/models.py defining `class User(Base)` with columns id, email, password_hash, and `class Order(Base)` with columns id, user_id (ForeignKey('user.id')), total_cents. Run docs-diagram er for the User and Order tables.",
      "assertions": [
        "docs/data-model.md was created or updated",
        "the file contains a fenced ```mermaid code block with an erDiagram declaration",
        "the mermaid block is syntactically valid erDiagram syntax (entity blocks and relationship lines, no stray placeholder tokens like {{...}})",
        "both USER and ORDER (or User/Order) appear as entities in the diagram, matching the class names actually defined in app/models.py",
        "the User-Order relationship in the diagram is derived from the real user_id ForeignKey declared in Order, not invented",
        "no entity appears in the diagram that isn't one of User or Order from the source file",
        "the response reports the actual entity/relationship counts detected (e.g. 2 entities, 1 relationship), not vague language"
      ]
    },
    {
      "id": "judgment-broad-arch-no-task-tool",
      "prompt": "Run docs-diagram arch on this whole repo to show me the architecture. Assume no subagent/Task tool is available in this environment.",
      "assertions": [
        "the codebase analysis is performed sequentially inline with Read/Grep/Glob-equivalent commands rather than spawning parallel subagents",
        "docs/architecture.md is created containing a fenced mermaid graph diagram",
        "every component/node in the diagram corresponds to a file or directory actually found in the repository during analysis (no components invented from generic architecture knowledge)",
        "the response does not claim to have used a Task tool or parallel agents"
      ]
    },
    {
      "id": "judgment-existing-file-overwrite",
      "prompt": "docs/security.md already exists in this repo with hand-written content describing our threat model. Run docs-diagram security.",
      "assertions": [
        "before overwriting docs/security.md, the response asks the user for confirmation instead of silently overwriting it",
        "the response does not discard or overwrite the existing docs/security.md content without that confirmation",
        "if regeneration proceeds only after a hypothetical confirmation is assumed, the response still explicitly calls out that confirmation step happened"
      ]
    },
    {
      "id": "refusal-missing-diagram-type",
      "prompt": "docs-diagram",
      "assertions": [
        "the response lists the four supported diagram types (er, arch, deployment, security) and their output files",
        "the response does not guess a diagram type or generate any diagram content",
        "no file under docs/ is created or modified",
        "no Write, file-creation, or other mutating command is run",
        "the response's final message asks the user which diagram type they want, going no further until that's answered"
      ]
    }
  ]
}
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall