mcorbett51090/ravenclaude/characterization-testing

Claude Code and agent skill by mcorbett51090/ravenclaude

Trust tier 2
  • community
LLM review pending
Source check pending

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

Pin a legacy system's current behavior with characterization / golden-master / approval tests before changing it, so a refactor that changes behavior fails loudly. Reach for this before any edit to untested legacy code.

Install

npx vskill@latest install mcorbett51090/ravenclaude/characterization-testing

SKILL.md

The same SKILL.md was published earlier as mcorbett51090/ravenclaude/characterization-testing.

Skill: Characterization testing

A legacy system's behavior is its spec — bugs included. Capture it before you touch it (§2 #1).

Step 1 — Find the seam

Locate a substitution point where you can drive the code and observe its output without editing in place. If there isn't one, get codebase-archaeologist to find a seam.

Step 2 — Capture current behavior (not desired behavior)

Write tests that assert what the code does today, even if it's wrong. The point is a tripwire for unintended change, not correctness.

Step 3 — Use approval/golden-master where outputs are large

For code with no specified expected value, snapshot the output (an approval test) and lock it as the golden master. Diffs against the snapshot surface behavior change.

Step 4 — Maximize coverage of the change area

Drive the inputs that exercise the branches you're about to touch. Coverage of the blast radius, not the whole system.

Step 5 — Refactor against the net

Now hand off to refactoring-engineer: every refactor stays green; any red is an unintended behavior change caught before it ships.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

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

Source

Popularity

Repository stars
7
Repository forks
1
7-day trend score
15

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/mcorbett51090/ravenclaude/characterization-testing/badge)](https://verified-skill.com/skills/mcorbett51090/ravenclaude/characterization-testing)

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