
AI System Inventory (EU AI Act)
Track each AI system's EU AI Act role and risk tier, and classify new systems step by step.
- Repository
- anthropics/claude-for-legal
- Stars
- 9,870
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This skill turns the agent into a structured record-keeper for AI systems under the EU AI Act. You can add a system through a short interview, list the inventory as a table with counts by risk tier, and view or edit individual records. It walks through two classification questions for each system: what role your organization plays (provider, deployer, importer, and so on) and which risk tier applies (prohibited, high-risk, limited, minimal, or general-purpose AI). The skill stores these records in a YAML file on the user's machine and flags that legal conclusions need a lawyer's verification against the current AI Act text.
Great for
- In-house legal or governance teams keeping an EU AI Act register of company AI tools
- Compliance staff who need a per-system record of role and risk tier with reasons
- Organizations that use several AI tools and need to track which ones touch EU people
Skip it if
- Final legal advice: the skill itself says article mappings are complex and must be verified by a lawyer against the current AI Act text
- Individuals with no EU nexus, since the EU-related questions do not apply to them
- Anyone expecting it to derive obligations automatically: it deliberately does not use a lookup table for obligations
Try asking
- “ai inventory”
- “add an ai system”
- “classify this ai system”
- “what systems do we have”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
ai-inventory.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip ai-inventory.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip ai-inventory.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Run the cold-start interview first so the CLAUDE.md config has no placeholder markers, or the skill will redirect you.
- Add every system even if classification is incomplete; you can run classify later for each sys-NNN ID.
- Treat the role_basis and tier_basis outputs as drafts and verify the cited AI Act articles before relying on them.
What's inside
- SKILL.md
Redistributed unmodified from anthropics/claude-for-legal under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.