Hey Product Hunt! agentools.sh is live there today. Tell me what to add →
Sponsor

Testing skills

11 Testing skills, most installed first. I read every file in each one to see what it can access, and show the security audits from skills.sh. See all skills.

Best for Solo founders who want their coding agent to test and debug before shipping

This page lists the Agent Skills that help a coding agent test, debug and audit your code, sorted by total installs on skills.sh. I read every skill file by file to rate its access (Low, Medium or High), and I show the skills.sh security audits next to each one.

Start with Test-Driven Development if you want strict red-green-refactor on every change. Pair it with Systematic Debugging, which makes the agent find the root cause before it edits code. Both come in the Superpowers pack from Jesse Vincent.

For a local web app, try Web App Testing from Anthropic, which drives your UI with Python Playwright. If you want a security pass, Security Audit (Cloudflare) runs six phases and has a second agent try to disprove every finding.

1agent-browservercel-labs/agent-browserFast browser automation CLI for agentsHigh1 warn1.2M2Systematic Debuggingobra/superpowersFind the root cause before touching codeMedium3/3 pass310K3Superpowersobra/superpowersFull TDD methodology pack for coding agentsMedium1 warn273K4Test-Driven Developmentobra/superpowersStrict red-green-refactor for every changeMedium3/3 pass262K5Web App Testinganthropics/skillsTest your local web app with PlaywrightMedium2 warn190K6Break UIemilkowalski/skillsStress-test UI with worst-case dataLow3/3 pass68.4K7EAS Simulatorexpo/skillsRun your app on a cloud iOS or Android simulatorHigh1 warn20.9K8Interface Reviewjakubkrehel/skillsReview a branch or PR for interface regressionsMedium1 warn19.2K9Break (Jakub Krehel)jakubkrehel/skillsStress test a component in every scenarioLow3/3 pass14.3K10Semgrep Scantrailofbits/skillsSemgrep security scan with approval gateHigh1 warn9.6K11Security Audit (Cloudflare)cloudflare/security-audit-skillA six-phase security audit where every finding is checked by a second agentMediumNo audits—

Questions

What is the best Agent Skill for browser testing with Claude Code?

agent-browser from Vercel is the most installed skill on this page. It is a thin stub that loads its full workflow at runtime, and its access is High. If you want something with Medium access, Web App Testing tests your local web app with Python Playwright scripts and captures screenshots and console logs.

How do I install the Test-Driven Development skill in Claude Code?

Run /plugin install superpowers@claude-plugins-official. That installs the Superpowers pack, which includes Test-Driven Development and Systematic Debugging. All three are MIT licensed.

What is the difference between Semgrep Scan and Cloudflare Security Audit?

Semgrep Scan from Trail of Bits runs Semgrep rulesets you approve and merges the results into SARIF. Security Audit (Cloudflare) maps your architecture and trust boundaries, then has a second agent try to disprove every finding before writing a JSON and readable report. Install it with npx skills add https://github.com/cloudflare/security-audit-skill --skill security-audit (add --global for a user-level install).

Which testing skills work with Cursor?

agent-browser, Systematic Debugging, Superpowers, Test-Driven Development and Security Audit (Cloudflare) all list Cursor support. Web App Testing lists only Claude Code, Claude.ai and the Claude API, and Semgrep Scan lists Claude Code and Codex.