project-mapper

Project Mapper

Scan project structure, detect stack, classify architecture. Auto-link to gap-analysis.

When to Use

New project, unfamiliar codebase, "how is this structured", onboarding, pre-refactor assessment.

Project Type Classification

Reference: gap-analysis skill (Phase 0.1) for classification tables. Avoid duplication.

Stack Detection Signals

go.mod→Go(gin/echo/fiber/chi) | package.json→Node(express/next/react/vue/nest) | pyproject.toml/Pipfile→Python(django/fastapi/flask) | Cargo.toml→Rust(actix/axum/rocket) | composer.json→PHP(laravel/symfony) | *.csproj→C#/.NET | build.gradle/pom.xml→JVM(spring-boot/quarkus) | Dockerfile→Docker | docker-compose.yml→Compose | makefile/justfile/Taskfile.yml→Build/Task Runner | turbo.json/nx.json→Monorepo | pnpm-workspace.yaml/yarn.lock→Workspaces.

Auto-Chain Protocol

After map: "Project classified as {tech layer}/{business type}. Run gap-analysis with {template}? (Y/n)". Only auto-execute on explicit confirm or --auto.

Rules

1. Start root ls/Get-ChildItem → top-level detection. 2. Classify from signals (multiple passes if ambiguous). 3. Drill 2-3 levels for architecture. 4. Validate: test count, coverage, lint, Docker size. 5. >50 files → summary counts per layer. 6. Adapt output to project size. 7. Suggest gap-analysis. 8. Save to engram topic_key: architecture/project-map:{name}.

Anti-Rationalization

Rationalization	Red Flag	Verification
"stack adivinado sin evidence"	Stack sin señales go.mod/package.json/pyproject/Cargo/composer	Verificar Stack Detection Signals con file:line + drill 2-3 niveles + test count/coverage/lint
"dependency map manual en vez de tool-backed"	Map manual sin validación tool-backed	Verificar dependency map tool-backed + Auto-Chain Protocol a gap-analysis file:line + engram architecture/project-map
"clasificación Tech/Biz sin checklist"	Clasificación sin tablas gap-analysis Phase 0.1	Verificar Project Type Classification vía gap-analysis tables + validación >50 files summary file:line

Red Flags

• Doing work without checking output format → STOP, re-read skill
• Second occurrence of same rationalization → force RED zone

Verification

• Output matches skill ## Output contract + file:line citaton
• cross-ref-check.ps1 → SKILL.md OK

Security Playbook Dispatch

API→auth-hardening | SPA→best-practices+auth-hardening | backend→data-quality | containers→container-security | IA→llm-security Monorepo/mixed: dispatch per package/target Hybrid single-repo (Next/Nuxt/Remix/SvelteKit): detect API routes (app/api, pages/api, server/api, src/server) + SPA -> apply BOTH playbooks (backend=data-quality + SPA=best-practices+auth-hardening), never one only

changelog: odd/tasks/mejora-security.md (2026-09-18, slice 6) changelog: odd/tasks/mejora-security.md (2026-09-18, slice 5)

Refs

gap-analysis · research · execution-mode · sdd · skill-graph · engram-protocol

Reference

Architecture Detection table + Output Format → docs/skills/project-mapper/reference.md