AI agent skills for Claude Code, Codex and Cursor.
One skill, one repeatable task. Cleaner commits, safer repos, stronger PRs and handoffs. Useful automation without the theatrical complexity.
Not a sprawling assistant. A sharp tool.
An agent skill is a small SKILL.md package that teaches your AI coding assistant one task. Each JustHandled Labs skill does a single job you can name: generate the commit message, catch the row-level-security hole, write the handoff before the context runs out. No vague "productivity," no kitchen-sink agent that does everything a little. Just the specific workflows that are easier to run, review, and repeat.
Start with the job you need handled
Or start with 11 free skills →Popular and proven
See all 98 skills →README Generator
FreeGenerates a README from your actual repo, not a hallucinated guess. Reads your manifests for real install steps and dependencies across Node, Python, Go, Rust, and Java.
Git Commit Writer
FreeGenerates a Conventional Commit message from your staged diff, detects type and scope from the changed files, and flags breaking changes without touching the repo.
AI Build Sanity Check
$13Find the mocks, placeholders, swallowed errors, and fake-success endpoints that make an AI-built app look finished when it is not.
Code Comment Cleaner
$5Find commented-out dead code across eight languages while preserving TODOs, license headers, disabled tests, and prose. Preview every removal before writing.
Lead Enricher
FreeTurn a company URL or name into a practical B2B lead profile: industry, size, tech stack, likely decision-makers, estimated email patterns, and buying signals.
Pipeline Reporter
FreeTurn a messy CRM CSV export into a clean weighted forecast by stage, rep-performance report, and close-date outlook.
PII & Data-Leak Scanner
$15Scan schemas, seed data, configuration, and logs for personal data before it leaks. Every finding includes the evidence and where it appeared.
Pipeline Reporter Pro
$19Turn CRM exports or live HubSpot, Salesforce, or Pipedrive data into a 12-month forecast with scenarios, win/loss themes, and an executive-ready report.
Skill Injection Scanner
$15Scan a SKILL.md package for prompt injection, secret exfiltration, and unsafe-command instructions before you install or publish it.
Who builds these
JustHandled Labs is built by H.J. Westerfield, with a background in communications, editing, project coordination, and practical AI systems. The catalog has grown well past developer workflow: repo and app safety, model risk and resilience for the era of models that get pulled or deprecated, pre-launch checks for non-technical founders, brand and social-presence tools, writing and anti-slop, and the agent-native layer (AGENTS.md, llms.txt, tool schemas). The throughline is the same as day one: one skill, one repeatable task, useful automation without theatrical complexity.
Questions
What is an agent skill?
A small package of instructions and scripts, written to the open SKILL.md standard, that teaches an AI coding assistant how to do one specific task. You add it once, and the assistant uses it when the work matches.
Which assistants do these work with?
Anything that reads SKILL.md, including Claude Code, Codex, and Cursor. The format is open, so the skills are not locked to a single tool or vendor.
Are they safe to run on my repo?
The scanning skills are read-only by default and preview every change before writing. Nothing touches your files or your database without your confirmation.
How do I install one?
Browse the full catalog here, open a skill, then buy or download it on Agensi. Each listing includes setup steps and what it reads.