What is grcskills?
grcskills is an open collection of practical GRC skills for AI agents, written around the steps experienced GRC practitioners follow. The skills work with Claude Code, Codex, Cursor and other agents that read SKILL.md. They cover tasks such as policy drafting, audit preparation, evidence, privacy and AI governance. An agent can use your repository, policy folder and accessible exports for a setup review, while you retain responsibility for judging findings and approving the work.
grcskills Features:
Skill Library: The collection includes 64 skills, each describing a process, its quality bar, the files it produces and where a person makes a decision.
GRC Work Areas: Skills are grouped across ten areas, including risk, audit preparation, evidence, vendors, privacy, incidents and AI systems.
Setup Review: A setup review selects ten skills that appear relevant to your setup and checks for issues, duplication or missing material in accessible files.
Written Deliverables: The setup review writes its shortlist, audit findings and draft 30-day plan to files, without marking any work as complete.
Framework Answers: An optional connection to ISMS Copilot lets an agent ask framework questions and receive answers with clause numbers for you to check.
Controlled Questions: When using ISMS Copilot, the skills direct the agent to show you a short question first and not send file contents, secrets, personal data or evidence.
grcskills Benefits:
Clear Starting Point: A ranked shortlist helps you focus on GRC tasks that appear relevant to your setup instead of sorting through every available skill.
Visible Gaps: Surfacing broken, duplicated or missing material gives you a clearer basis for deciding what needs attention.
Reviewable Work: Deliverables saved to files make it easier to examine the agent's proposed findings and plan before deciding what to accept.
Human Oversight: Keeping decisions and sign-off with you helps prevent draft GRC work from being treated as approved simply because an agent produced it.
More Specific Answers: Clause-numbered framework answers give you a concrete reference to verify rather than relying only on an agent's recollection.
Use Cases:
Access Review Prep: Prepare an access review by reconciling account lists with HR, flagging leavers and administrators, and creating sheets for managers.
AI Policy Drafting: Develop acceptable-use rules that identify approved AI tools, what data staff may enter and how agents and connectors should be used.
EU AI Act Review: Classify an AI system by assessing scope, provider or deployer role, prohibited practices and high-risk tests, with reasoning to review.
AI Impact Assessment: Assess who an AI system could affect, how it could fail them, what safeguards exist and what still needs sign-off.
Draft Audit Check: Review an AI-drafted GRC artefact by tracing samples to their sources, testing what it did not flag and completing a review record.

