MCP HubMCP Hub
Volver a habilidades

tidy-project-structure

pjt222
Actualizado Yesterday
3 vistas
17
2
17
Ver en GitHub
Desarrollogeneral

Acerca de

Esta habilidad de Claude organiza y limpia la estructura del proyecto moviendo archivos a directorios convencionales, actualizando READMEs obsoletos y consolidando archivos de configuración. Es útil cuando los proyectos tienen archivos dispersos, documentación desactualizada o desviación en la configuración entre entornos. La herramienta se centra en mejoras estructurales sin alterar la lógica central del código.

Instalación rápida

Claude Code

Recomendado
Principal
npx skills add pjt222/agent-almanac -a claude-code
Comando PluginAlternativo
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternativo
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/tidy-project-structure

Copia y pega este comando en Claude Code para instalar esta habilidad

Documentación

tidy-project-structure

Use When

Project org drifted from conventions:

  • Files scattered, no clear org
  • READMEs outdated | broken examples
  • Config files multiplied (dev, staging, prod drift)
  • Deprecated in project root
  • Naming inconsistent across dirs

Do NOT use for code refactoring | dep restructuring. Focus = file org + doc hygiene.

In

ParameterTypeRequiredDescription
project_pathstringYesAbsolute path to project root
conventionsstringNoPath to style guide (e.g., docs/conventions.md)
archive_modeenumNomove (default) or delete for deprecated files
readme_updatebooleanNoUpdate stale READMEs (default: true)

Do

Step 1: Audit Layout

Cmp current structure vs conventions | language best practices.

Common conventions by lang:

JavaScript/TypeScript:

src/          # Source code
tests/        # Test files
dist/         # Build output (gitignored)
docs/         # Documentation
.github/      # CI/CD workflows

Python:

package_name/      # Package code
tests/             # Test suite
docs/              # Sphinx docs
scripts/           # Utility scripts

R:

R/                 # R source
tests/testthat/    # Test suite
man/               # Documentation (generated)
vignettes/         # Long-form guides
inst/              # Installed files
data/              # Package data

Rust:

src/          # Source code
tests/        # Integration tests
benches/      # Benchmarks
examples/     # Usage examples

Got: List of files/dirs violating saved to structure_audit.txt.

If err: No conventions doc'd → use language-std defaults.

Step 2: Move Misplaced Files

Relocate to conventional dirs.

Common moves:

  1. Test files outside tests/tests/
  2. Docs outside docs/docs/
  3. Build artifacts in src/ → delete (gitignored)
  4. Config in root → config/ | .config/

Per move:

# Check if file is referenced anywhere
grep -r "filename" .

# If no references or only relative path references:
mkdir -p target_directory/
git mv source/file target_directory/file

# Update any imports/requires
# (language-specific — see repair-broken-references skill)

Got: All files in conventional locations; git history preserved via git mv.

If err: Moving breaks imports → update import paths | escalate.

Step 3: README Freshness

ID stale info in all READMEs.

Staleness indicators:

  1. Last mod >6 mo ago
  2. Old ver # references
  3. Broken links | code examples
  4. Missing sections (Install, Usage, Contributing)
  5. No license badge | broken badge links
# Find all READMEs
find . -name "README.md" -o -name "readme.md"

# For each README:
# - Check last modified date
git log -1 --format="%ci" README.md

# - Check for broken links
markdown-link-check README.md

# - Verify example code still runs (sample first example)

Got: List of stale READMEs in readme_freshness.txt w/ specific issues.

If err: markdown-link-check unavail → manually review external links.

Step 4: Update Stale READMEs

Fix broken links, update examples, add missing sections.

Std fixes:

  1. Replace broken badge URLs
  2. Update vers in install instructions
  3. Fix broken example code (run to verify)
  4. Add missing sections (template from conventions)
  5. Update copyright year

README template:

# Project Name

Brief description (1-2 sentences).

## Installation

```bash
# Language-specific install command

Usage

# Basic example

Documentation

Link to full docs.

Contributing

Link to CONTRIBUTING.md or inline guidelines.

License

LICENSE badge and link.


**Got:** All READMEs updated; examples verified to run.

**If err:** Example code can't be verified → mark w/ warning comment.

### Step 5: Review Config Files

ID drift + consolidate duplicate settings.

**Common config issues**:
1. Multiple `.env` (`.env`, `.env.local`, `.env.dev`, `.env.prod`)
2. Duplicate settings across configs
3. Hardcoded secrets (use env vars)
4. Outdated API endpoints | feature flags

```bash
# Find all config files
find . -name "*.config.*" -o -name ".env*" -o -name "*.yml" -o -name "*.yaml"

# For each config:
# - Check for duplicate keys
# - Grep for hardcoded secrets (API keys, tokens, passwords)
grep -E "(api[_-]?key|token|password|secret)" config_file

# - Compare dev vs prod settings
diff .env.dev .env.prod

Got: Config drift doc'd in config_review.txt; secrets flagged for escalation.

If err: Diff shows major divergence → escalate to devops-engineer.

Step 6: Archive Deprecated Files

Move | delete files no longer needed.

Candidates:

  • Commented-out configs (nginx.conf.old)
  • Legacy scripts not run in >1y
  • Backup files (file.bak, file~)
  • Build artifacts accidentally committed

Archive process:

# Create archive directory (if archive_mode=move)
mkdir -p archive/YYYY-MM-DD/

# For each deprecated file:
# 1. Verify not referenced anywhere
grep -r "filename" .

# 2. Check git history for last modification
git log -1 --format="%ci" filename

# 3. If not modified in >1 year and no references:
if [ "$archive_mode" = "move" ]; then
  git mv filename archive/YYYY-MM-DD/
else
  git rm filename
fi

# 4. Document in ARCHIVE_LOG.md
echo "- filename (reason, last modified: DATE)" >> ARCHIVE_LOG.md

Got: Deprecated archived; ARCHIVE_LOG.md updated.

If err: Uncertain if deprecated → leave + doc in report.

Step 7: Verify Naming Conventions

Check inconsistent file naming across project.

Common conventions:

  • kebab-case: my-file.js (JS/web)
  • snake_case: my_file.py (Python)
  • PascalCase: MyComponent.tsx (React)
  • camelCase: myUtility.js (JS fns)
# Find files violating conventions
# Example: Python project expecting snake_case
find . -name "*.py" | grep -v "__pycache__" | grep -E "[A-Z-]"

# For each violation, either:
# 1. Rename to match conventions
# 2. Document exception (e.g., Django settings.py convention)

Got: All files follow conventions | exceptions doc'd.

If err: Renaming breaks imports → update references | escalate.

Step 8: Generate Tidying Report

Doc all structural changes.

# Project Structure Tidying Report

**Date**: YYYY-MM-DD
**Project**: <project_name>

## Directory Changes

- Moved X files to conventional directories
- Created Y new directories
- Archived Z deprecated files

## README Updates

- Updated W stale READMEs
- Fixed X broken links
- Verified Y code examples

## Config Cleanup

- Consolidated X duplicate settings
- Flagged Y hardcoded secrets for removal
- Documented Z config drift issues

## Files Archived

See ARCHIVE_LOG.md for full list (Z files).

## Naming Convention Fixes

- Renamed X files to match conventions
- Documented Y exceptions

## Escalations

- [Config drift requiring devops review]
- [Hardcoded secrets requiring security audit]

Got: Report saved to TIDYING_REPORT.md.

If err: (N/A — generate regardless)

Check

Post-tidy:

  • All files in conventional dirs
  • No broken links any README
  • README examples verified
  • Config files reviewed for secrets
  • Deprecated archived w/ docs
  • Naming conventions consistent
  • Git history preserved (git mv not mv)
  • Tests still pass after moves

Traps

  1. Break Relative Imports: Moving breaks relative paths. Update refs | use absolute.
  2. Lose Git History: mv not git mv → loses history. Always git cmds for moves.
  3. Over-Organize: Too many nested dirs → harder navigation. Flat until complexity demands.
  4. Delete vs Archive: Direct delete → no recovery. Always archive first unless certain.
  5. Ignore Language Conventions: Personal pref over language std. Follow established.
  6. Not Updating Docs: Moving w/o updating README paths → broken docs.

Repositorio GitHub

pjt222/agent-almanac
Ruta: i18n/caveman-ultra/skills/tidy-project-structure
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Habilidades relacionadas

qmd

Desarrollo

qmd es una herramienta CLI de búsqueda e indexación local que permite a los desarrolladores indexar y buscar en archivos locales mediante búsqueda híbrida que combina BM25, embeddings vectoriales y reranking. Es compatible tanto con uso desde la línea de comandos como con modo MCP (Model Context Protocol) para integración con Claude. La herramienta utiliza Ollama para los embeddings y almacena los índices localmente, lo que la hace ideal para buscar documentación o bases de código directamente desde la terminal.

Ver habilidad

subagent-driven-development

Desarrollo

Esta habilidad ejecuta planes de implementación asignando un nuevo subagente para cada tarea independiente, con revisión de código entre tareas. Permite una iteración rápida mientras mantiene controles de calidad a través de este proceso de revisión. Úsala cuando trabajes en tareas mayormente independientes dentro de la misma sesión para garantizar un progreso continuo con verificaciones de calidad integradas.

Ver habilidad

mcporter

Desarrollo

La habilidad mcporter permite a los desarrolladores gestionar y llamar servidores del Protocolo de Contexto de Modelo (MCP) directamente desde Claude. Proporciona comandos para listar servidores disponibles, llamar a sus herramientas con argumentos, y manejar la autenticación y el ciclo de vida del daemon. Utiliza esta habilidad para integrar y probar la funcionalidad de servidores MCP en tu flujo de trabajo de desarrollo.

Ver habilidad

adk-deployment-specialist

Desarrollo

Esta habilidad despliega y orquesta agentes Vertex AI ADK utilizando el protocolo A2A, gestionando el descubrimiento de AgentCard, el envío de tareas y soportando herramientas como el Sandbox de Ejecución de Código y el Banco de Memoria. Permite construir sistemas multiagente con patrones de orquestación secuencial, paralela o en bucle en Python, Java o Go. Úsela cuando se le solicite desplegar agentes ADK u orquestar flujos de trabajo de agentes en Google Cloud.

Ver habilidad