About AI Skills Hub

An internal registry for AI skills at our company.

AI Skills Hub is our internal directory for AI skills — reusable agent capabilities that any teammate can discover, install, and build on. Think of it as an inner-source package registry designed around how our team actually works: short review cycles, company-wide visibility, and guardrails tailored to our security and compliance requirements.

Anyone on a verified company email can browse. Publishing and installing follow a lightweight review flow so that skills stay trustworthy as the catalogue grows.

What’s allowed

Agent tooling that boosts engineering productivity (testing, code review, refactoring).
Documentation, changelog, and knowledge-management helpers.
Developer experience utilities (linting rules, project scaffolding, CI integrations).
Internal-process automations (ticket triage, meeting summarisation, release notes).
Data analysis, reporting, and visualisation skills that respect company data policies.

What’s not allowed

Skills that bypass authentication, authorisation, or rate limits on any system.
Platform abuse — scraping, spam, botting, or circumventing vendor TOS.
Fraud, deception, or social engineering (including phishing content generators).
Privacy invasion — surveillance of employees or customers, doxxing, or PII harvesting.
Non-consensual impersonation of real people, including coworkers and customers.
Explicit sexual or violent content, or skills that remove safety guardrails.
Hidden execution, obfuscated payloads, or skills that fetch and run remote code.
Anything that violates company security, legal, or HR policies.

How moderation works

Every skill starts as a draft. When the author submits it for review, an automated security scan runs and the skill moves into the reviewer queue. Reviewers (tagged in-app) check the skill’s description, instructions, and content against the rules above, then approve, request changes, or reject it.

Approved skills are publicly listed across AI Skills Hub. If you believe an approved skill violates these rules, open a report on the skill detail page — admins can suspend a skill at any time while the issue is investigated.

Questions or feedback? Reach out in our internal #clawhub channel.