ludo-technologies/pyscn/cli-analysis

Claude Code and agent skill by ludo-technologies/pyscn

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Run the pyscn command-line tool for Python code quality analysis - CI/CD quality gates, HTML/JSON/CSV reports, full analysis runs, and project configuration. Use when user wants a CI check, a shareable report file, or to configure pyscn for a project.

Install

npx vskill@latest install ludo-technologies/pyscn/cli-analysis

SKILL.md

Python Code Quality Analysis with the pyscn CLI

Use the pyscn command-line tool when the task needs report files, CI integration, or project configuration.

No install needed: uvx pyscn@latest <command> (or pip install pyscn for a permanent install).

Commands

CommandPurpose
pyscn analyze <path>Comprehensive analysis: complexity, dead code, clones, coupling (CBO), cohesion (LCOM), dependencies, architecture
pyscn check <path>Fast pass/fail quality gate for CI with predefined thresholds
pyscn initGenerate a .pyscn.toml config file

analyze — full analysis and reports

bash
pyscn analyze . # human-readable summary + health score
pyscn analyze --json src/ # also write a JSON report file
pyscn analyze --html src/ # interactive HTML report
pyscn analyze --select complexity,deadcode src/ # only specific analyses
pyscn analyze --skip-clones src/ # clones are the slowest analysis
pyscn analyze --min-complexity 10 --min-severity critical src/

Key flags:

  • --json / --html / --csv / --yaml: generate a report file
  • --select: complexity,deadcode,clones,cbo,lcom,deps,communities
  • --skip-clones, --skip-cbo, --skip-deps, ...: skip individual analyses
  • --clone-threshold (default 0.65), --min-complexity (default 5), --min-cbo
  • --no-open: don't auto-open HTML reports in a browser (use in scripts)

Report files are NOT written to stdout. They go to .pyscn/reports/ under the current working directory with timestamped names like analyze_20260702_221139.json, and the path is printed on completion. To consume results programmatically: run with --json, capture the printed path (or take the newest file in .pyscn/reports/), then read that file.

check — CI quality gate

bash
pyscn check . # complexity + deadcode + clones
pyscn check --select complexity --max-complexity 10 src/
pyscn check --select deps src/ # fail on circular dependencies
pyscn check --allow-dead-code --skip-clones src/
pyscn check -q . # quiet: output only on issues

Exit codes: 0 no issues, 1 quality issues found, 2 analysis failed (invalid input, missing files). Default gates: complexity > 10 fails, critical dead code fails, clones warn only, dependency cycles fail.

--select accepts: complexity, deadcode, clones, deps, mockdata, di (dependency-injection anti-patterns).

Configuration

Settings load in priority order: command-line flags > .pyscn.toml > [tool.pyscn] in pyproject.toml > defaults. Run pyscn init to scaffold .pyscn.toml when a project wants persistent thresholds.

Reporting Results

Summarize the health score and grade, list the specific functions/files behind each failing category, and suggest fixes. For CI setup, recommend pyscn check in the pipeline and pyscn analyze --html for periodic deep reviews.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
1,029
Repository forks
69
7-day trend score
35

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/ludo-technologies/pyscn/cli-analysis/badge)](https://verified-skill.com/skills/ludo-technologies/pyscn/cli-analysis)

Something wrong with this skill or its page? Tell us.