test-me
Decide whether a tool really does the job, from evidence of using it, not from its README.
Test the tool before you build on it.
For a new app, library or agent tool, or a feature you plan to copy from one.
test-me has your AI agent test the tool the way a careful reviewer would. Each promise becomes a single, testable claim. The test is fixed before the tool runs, and the tool is used on synthetic data.
Every claim gets a plain verdict: keep the current method, run a limited pilot, adopt it, or drop it for now. "Keep what you have" is often the right answer, and the skill says so.
Two commands in Claude Code.
Any agent that reads SKILL.md folders can use the skill folder directly.
/plugin marketplace add 0xOzen/test.me /plugin install test-me@aequai-skills
One page, decision first.
For founders, operators, developers, and teams that handle sensitive data.
A one-page report with the decision first, the evidence for each claim, and what was not checked. MIT licence.