| name | refactoring |
| description | Find refactoring targets in Python code using pyscn - duplicate code (clones), overly complex functions, and dead code. Use when user asks about refactoring, code duplication, complexity hotspots, unreachable code, or cleaning up a codebase. |
Python Refactoring Analysis with pyscn
Run the pyscn CLI to locate concrete refactoring targets. No install needed: uvx pyscn@latest <command>.
Commands
| User Request | Command |
|---|
| "Find complex functions" | uvx pyscn@latest analyze --select complexity <path> |
| "Find duplicate code" | uvx pyscn@latest analyze --select clones <path> |
| "Find dead code" | uvx pyscn@latest analyze --select deadcode <path> |
| "What should I refactor first?" | uvx pyscn@latest analyze --select complexity,deadcode,clones <path> |
Key Flags
--min-complexity <n>: minimum complexity to report (default: 5). Risk levels: Low (<10), Medium (10-20), High (>20)
--clone-threshold <0.0-1.0>: minimum similarity for clone detection (default: 0.65)
--min-severity <critical|warning|info>: dead code severity floor (default: warning). Critical means code after return/break/continue/raise that can never execute.
--json: write a detailed report with line-level findings. Report files are NOT written to stdout; they go to .pyscn/reports/ and the path is printed on completion.
Prioritizing Findings
- Critical dead code: safe deletions, do these first.
- High-complexity functions (>20): extract functions, flatten conditionals.
- Clone groups spanning multiple files: extract shared helpers; clones within one file are usually quicker wins.
When suggesting a refactor, cite the specific function names, files, and line ranges from the results, and re-run the same command afterward to confirm the improvement.
MCP Alternative
If the pyscn-mcp MCP server is connected, you can use its tools instead of the CLI: check_complexity (min_complexity), detect_clones (similarity_threshold, min_lines), find_dead_code, or analyze_code with analyses: ["complexity", "clone", "dead_code"] to run several at once. Results are returned inline with no report files.