Drydock¶
An agentic toolchest for deconstructing, analyzing, and reconstructing software architectures.
-
MCP Server Included
Plug into Claude Code, Cursor, or VS Code. Your AI assistant gets architecture analysis as native tool calls.
-
Pure Python, Zero Deps
Every analyzer is stdlib-only. Clone and run — no
pip install, no virtual environment, no build step. -
Boundary Detection
One command tells you exactly which components can be safely extracted with zero dependency risk.
-
JSON-First Output
All tools default to structured JSON for AI consumption. Add
--markdownfor human-readable reports.
What It Does¶
Drydock analyzes codebases and produces structured intelligence that AI agents (or humans) can act on:
# One command — understand an entire project
python Tools/analyzers/context_compiler.py /path/to/project -o context.json
{
"summary": {
"total_source_files": 8,
"clusters_found": 1,
"bridge_files": 0
},
"extraction_suggestions": [
{
"name": "core",
"files": 6,
"extraction_risk": "low",
"reason": "No external dependencies - clean extraction"
}
]
}
Language Support¶
| Language | Code Map | Interfaces | Boundaries | Dependencies |
|---|---|---|---|---|
| Python | Full AST | Full AST | imports | AST + stdlib |
| JS/TS | Regex | Regex | imports | package.json |
| Rust | Regex | Regex | use/mod | Cargo.toml |
| Go | Regex | Regex | imports | go.mod |
| C# | Regex | Partial | — | — |
| Java | Regex | Partial | — | — |
Quick Links¶
- MCP Server Setup — Use from Claude Code / Cursor / VS Code
- CLI Quick Start — Run from the terminal
- All Tools — Complete tool reference
- API Reference — Auto-generated from docstrings