shakedownlisted

# Shakedown Comprehensive review of any project. Discovers the structure dynamically, reads everything, challenges every layer, finds bottlenecks, blind spots, and waste, then produces ranked recommendations you can implement in the same session. Works on: multi-agent platforms (Paperclip, CrewAI, AutoGen), data pipelines, web applications, CLI tools, monorepos, microservices - any project with files to read and architecture to challenge. --- ## Phase 1: Discover Project Structure Before reading anything, map the project. Do NOT assume folder names, frameworks, or conventions - discover them. ### 1a. Identify What Kind of Project This Is List files in the working directory and its immediate subdirectories to understand the project layout. Use your platform's tools (Glob, LS, or shell commands) to discover: ```bash # Example commands (adapt to your platform): ls -la find . -maxdepth 2 -type f | head -50 ``` Check for framework indicators by looking for common config files: `package.json`, `pyproject.toml`, `Cargo.toml`, `go.mod`, `Makefile`, `docker-compose.yml`, `.env`, `*.config.*` Check for agent/pipeline platforms by looking for platform-specific directories (`.paperclip`, `.crew`, `.autogen`, `.langchain`, etc.) Check for **agent skill indicators**: - `SKILL.md` in the project root -> this project IS an agent skill - `.agents/skills/` directory -> project contains or uses skills - Skill-like YAML frontmatter (`name:`, `description:` fields) - References to skil