Kaidera Skills Marketplace

Open-source skills for AI agents — build, share, and collaborate.

Kaidera Skills Marketplace is a community-driven repository of vetted, reusable skills that extend AI agent capabilities on the Kaidera Platform. Skills are modular instruction sets that teach agents new abilities - from code review to infrastructure deployment.

---

What Is a Skill?

A skill is a structured YAML file (skill.yaml or SKILL.md) that defines:

• What the agent can do (description, capabilities)
• How it does it (prompt additions, code, tool bindings)
• When it activates (workflow phases, triggers)
• Who created it (attribution, license)

Skills come in three types:

Type	Purpose	Example
Steering	Shapes agent behavior via prompt injection	code-review.SKILL.md
Tool	Executable Python code in a sandbox	api-test.SKILL.md
Hybrid	Both steering and executable code	tdd-workflow.SKILL.md

Skill Categories

Category	Description	Count
context/	Project and workspace awareness	6
development/	Code writing, review, testing	7
devops/	Deployment, infrastructure, CI/CD	5
security/	Auditing, scanning, incident response	5

How It Works

┌─────────────────────────────────────────────────────────────┐
│                    SKILLS LIFECYCLE                           │
│                                                              │
│  1. CONTRIBUTE  ──  Author writes a skill.yaml               │
│  2. SUBMIT      ──  PR to this repo (or sync via platform)   │
│  3. VET         ──  Automated security scan + sandbox test    │
│  4. REVIEW      ──  Platform admin reviews in Pending tab     │
│  5. APPROVE     ──  Skill moves to Approved, available to     │
│                     agents in the workbench                   │
│  6. ATTRIBUTE   ──  Original author credited (CC-BY-4.0)     │
│                                                              │
└─────────────────────────────────────────────────────────────┘

For Platform Users

1. Open the Skills Management page in the Kaidera admin panel
2. Click Add Skills and enter this repo URL (or any public GitHub repo with skills)
3. Skills are synced to the Pending tab for review
4. Review vetting reports, edit if needed, then Approve or Reject
5. Approved skills appear in the Workbench for agents to use

For External Repos

You can sync skills from any public GitHub repository — not just this one. For example:

https://github.com/coreyhaines31/marketingskills

The platform will walk the repo, find all skill.yaml / SKILL.md files, parse them, and import them to your Pending queue. The original author is always credited via the attribution fields.

Skill Format

Every skill follows the SKILL_FORMAT.md specification. Here's a minimal example:

---
name: code-review
version: 1.0.0
description: Reviews code for bugs, security issues, and best practices
author: Kaidera Team
skill_type: steering
category: development
trust_tier: official
tags: [code-review, quality, security]
default_permission: allow
---

## Instructions

When asked to review code, follow these steps:
1. Check for security vulnerabilities (OWASP Top 10)
2. Identify logic errors and edge cases
3. Suggest improvements for readability and performance
4. Verify test coverage for changed code

Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

Ways to contribute:

• Submit new skills via pull request
• Improve existing skills (bug fixes, better prompts)
• Report issues with skills that don't work as expected
• Suggest new skill categories

All contributions are licensed under CC-BY-4.0 — you retain credit as the original author, and the community can build on your work.

Attribution

This project uses the Creative Commons Attribution 4.0 International license. When we pull skills from external repositories, we:

1. Credit the original author in the skill's attribution_author field
2. Link back to the source repository in attribution_url
3. Preserve any additional license notes in attribution_notes

If you see your work here without proper credit, please open an issue and we'll fix it immediately.

Security

Every skill goes through a multi-gate security pipeline before it reaches an agent:

1. Static Analysis — Pattern matching for prompt injection, credential leaks, dangerous imports
2. Sandbox Testing — Isolated execution with restricted syscalls and network
3. Expert Review — Human approval required for community-submitted skills
4. Runtime Envelope — Hermetic XML boundary prevents skills from overriding system prompts

See SKILL_FORMAT.md for the full vetting specification.

License

This repository is licensed under CC-BY-4.0 (Creative Commons Attribution 4.0 International).

You are free to:

• Share — copy and redistribute skills in any format
• Adapt — remix, transform, and build upon skills for any purpose

Under the following terms:

• Attribution — You must give appropriate credit to the original author

---

Kaidera - The Machine That Builds Machines

kaidera.ai | Platform Repo