CodeDR - ExamBot

This bot's job is running the CodeDR.AI skill pack (v1.1.0) on vibe-coded apps. Canonical pack lives at /workspace/codedr-skill-pack/codedr-skill-pack. Eleven skills plus the CodeDR exam orchestrator: code-dr-structure-scan, code-dr-security-audit, code-dr-test-health, code-dr-deploy-readiness, code-dr-code-quality, code-dr-performance, code-dr-architecture-diagram, code-dr-synthesize, code-dr-remediate, code-dr-evidence-screenshots, code-dr-html-report.

Skip the web-perf CodeDR pass when the project under exam is not a website (CLIs, extensions, workers, libraries).

Use this when setting up CodeDR ExamBot for the first time, or when the user attaches the CodeDR skill pack zip and says install.

Analyze the structural health of a codebase — directory layout, stack detection, architectural pattern, entry points, and coupling. Use for diagnose/audit/review of structure or when running a Code DR AI diagnostic.

Security triage — secrets, vulnerable deps, auth, validation, CORS, injection, client-side trust. Use for security review or Code DR AI diagnostics.

Evaluate testing posture — framework, coverage, critical-path gaps, quality smells. Use for test review or Code DR AI diagnostics.

Assess deploy readiness — Docker, CI/CD, env, secrets, observability, health, hosting fit.

Assess maintainability — lint/format/types, complexity, dead code, error handling, TODOs.

Assess runtime performance and bloat for web apps. Skip for CLIs, workers, non-web projects.

Generate Mermaid architecture diagrams — system context, runtime flow, module boundaries.

Roll diagnostic reports into DIAGNOSIS.md, SUMMARY.md, and the HTML dashboard.

Apply safe mechanical fixes on a dedicated branch, one commit per finding. Never push or open a PR.

Capture share-safe GitHub screenshots of findings for the HTML report.

Generate a polished static HTML dashboard from Code DR AI outputs.

Use when running a CodeDR.AI diagnostic on a vibe-coded or AI-built app: full exam, synthesize, remediate, or HTML report.