Your Agent Leaks Secrets.
Ours Catch Them.

SKILL.md files from 8 agents running 24/7. 47 secrets caught pre-commit. 200+ cron failures diagnosed. Zero context blowouts in 6 months. Drop one file. Done.

Browse Skills →

Skills

Drop a file. Gain a capability. No build steps, no config files, no prayer.

Cron Doctor

FREE

Cron fails at 3am. You find out at 3:01 — not 9am.

That job that fails every third Tuesday? Caught. The cascade that starts with one timeout? Traced. Severity-ranked so you fix what's burning, ignore what can wait.

  • Pattern detection humans miss — intermittent ≠ ignorable
  • Severity triage: "wake up" vs. "Monday morning"
  • Root cause, not 500 lines of stack trace
monitoringcrondiagnosis
Try Free on Agent37

Self Monitor

FREE

Your agent fixes itself. Pages you only when it can't.

Context window at 80%? Compacted. Disk at 90%? Cleared. Service down? Restarted. You get pinged when it's actually beyond self-repair — which is rare.

  • Context window awareness — no more mid-task blowouts
  • Auto-remediation: disk, cache, services, memory
  • Escalates only what it can't handle alone
monitoringhealthinfra
Try Free on Agent37

Security Suite

$9.90

Catches the AWS key before git push does.

Five layers, always on. PII scanner flags 5 categories grep misses. Gitleaks stops secrets at commit. Prompt injection defense blocks malicious input before it hits your agent. One leaked key costs more than this.

  • PII: emails, phones, SSNs, API keys, credentials — all flagged
  • Pre-commit secret scanning (Gitleaks built in)
  • Prompt injection defense — external input is never trusted
securityauditPII

Structured Memory

$9.90

Instant recall. No context blowouts.

Three tiers: facts (permanent), episodes (session), scratchpad (working). Sub-100ms lookups. Auto-compaction kills stale entries before they eat your context window. Your agent remembers what matters, forgets what doesn't.

  • Facts persist. Session fades. Scratchpad clears. By design.
  • Sub-100ms retrieval — instant, not grep-and-pray
  • Auto-compaction keeps context lean and fast
memoryknowledgeorganization

Planning & Execution

$9.90

Ships features. Not "I refactored everything while you were out."

2-5 minute tasks. Checkpoints before drift becomes disaster. TDD-first, YAGNI-enforced. Your agent stays on target instead of rewriting half the codebase unprompted.

  • Small tasks = fewer rabbit holes = actual shipping
  • Checkpoints catch scope creep before it metastasizes
  • TDD-first: red, green, refactor — no exceptions
planningexecutionmethodology

🎁 The Full Stack

What 90% of buyers choose.

Security + Memory + Execution — everything your agent needs to run unsupervised. Most people grab all three anyway. This way you save $4.80 and skip the "should I get the other one?" decision.

$24.90 most popular
Get the Bundle

Why These Skills?

🏭

Battle-tested. Not weekend-project-tested.

8 agents. 6 months. 200+ cron jobs diagnosed. 47 secrets caught before commit. The memory bloat at 3am? Fixed. The race condition when two jobs overlap? Gone. We broke it so you don't have to.

🔌

Plain text. Zero lock-in.

SKILL.md is markdown. No proprietary format. No vendor dependency. Switch from Claude Code to Cursor to Codex tomorrow — your skills come with you. The file is yours.

🛡️

Security-first. Not security-later.

Built by a cat who reads OWASP for fun. External content is never trusted. Secrets never hit logs. Prompt injection defense is baked in from day one — not bolted on after the breach.

🔧

We eat our own dogfood

When these break, our agents break. We notice in hours, not quarters. Updates ship after a week of production testing on our own infrastructure. If it's not good enough for us, it doesn't ship.

Compatible with

Claude Code Codex CLI Cursor Windsurf OpenClaw

Frequently Asked Questions

What platforms do these work on?

+

Claude Code, Codex CLI, Cursor, Windsurf, OpenClaw — anything that reads markdown. It's a text file. If your agent can read, it can use these.

What's the SKILL.md format?

+

A markdown file your agent reads to learn new tricks. No npm install. No build step. No dependencies. Drop it in your workspace. That's literally it.

Can I try before buying?

+

Cron Doctor and Self Monitor are free — run them in production, see how it works. Premium skills: no-questions refunds if they're not for you. We're not going to fight you over $9.90.

How do updates work?

+

Buy once, own it. When we improve something (which happens — these run our own agents), re-download from Gumroad. No subscriptions. No expiry emails. No "please renew" nonsense.

One file. Tonight. Done.

Start free — Cron Doctor and Self Monitor cost nothing. Most people grab the bundle when they see how it works. Security + Memory + Execution for less than lunch.