---
id: eu-ai-act-high-risk-triage
track: coding
version: 0.0.4
label: warns
keyword: warn
where: write turn-end
sha256: d40396513324363c19515b8d83f69114e88f7480188eb90d8842cc0e780d6303
source: https://github.com/open-coder-ai/chock-catalog/tree/1e60d211c1bb767152c6ab0167932da06ab3a10f/compliance/eu-ai-act-high-risk-triage
catalog_ref: 1e60d211c1bb767152c6ab0167932da06ab3a10f
---

# EU AI Act — High-Risk Triage

eu-ai-act-high-risk-triage · version 0.0.4

Warns — Warns on an agent's file writes and at turn end. Chock adds friction. It is not a security boundary. 

In a git repo: advisory (The agent reads the text, and nothing stops a violation.)



## What it does

Warns when code puts an AI system into an EU AI Act Annex III high-risk domain — biometrics, critical infrastructure, education, employment, essential services and credit, law enforcement, migration, justice and elections — and asks for an owner of the Article 9-15 obligations before the capability ships. Use when adding scoring, ranking, eligibility, or screening over people. Do NOT use for banned practices (see eu-ai-act-prohibited-practices) or for systems with no natural-person impact.

## Its label on each agent

- Claude Code (tested): Warns — Warns on an agent's file writes and at turn end. Builder: /coding/builder/#client=claude-code&p=eu-ai-act-high-risk-triage
- Cursor (tested): Warns — Warns on an agent's file writes and at turn end. Builder: /coding/builder/#client=cursor&p=eu-ai-act-high-risk-triage
- Codex (tested): Warns — Warns on an agent's file writes and at turn end. If the hook errors, the action goes through (not witnessed; the vendor's row is only partly live-run). Builder: /coding/builder/#client=codex&p=eu-ai-act-high-risk-triage
- GitHub Copilot (tested): Warns — Warns at turn end only. It does not check a file write when it happens. If the hook errors, the action goes through (not witnessed; the vendor's row is only partly live-run). Builder: /coding/builder/#client=copilot&p=eu-ai-act-high-risk-triage
- Devin (tested): Warns — Warns at turn end only. It does not check a file write when it happens. If the hook errors, the action goes through (vendor docs only, not witnessed). Builder: /coding/builder/#client=devin&p=eu-ai-act-high-risk-triage
- Grok (tested): Not graded yet for Grok. Builder: /coding/builder/#client=grok&p=eu-ai-act-high-risk-triage
- goose (tested): Not graded yet for goose. Builder: /coding/builder/#client=goose&p=eu-ai-act-high-risk-triage
- Antigravity (tested): Not graded yet for Antigravity. Builder: /coding/builder/#client=antigravity&p=eu-ai-act-high-risk-triage
- OpenHands (untested): Not graded yet for OpenHands. Builder: /coding/builder/#client=openhands&p=eu-ai-act-high-risk-triage
- Kimi Code (untested): Not graded yet for Kimi Code. Builder: /coding/builder/#client=kimi-code&p=eu-ai-act-high-risk-triage
- Factory Droid (untested): Not graded yet for Factory Droid. Builder: /coding/builder/#client=factory-droid&p=eu-ai-act-high-risk-triage

## What it misses

This policy does not state what it misses.

## Examples

- Warns: At write time the agent adds a resume ranker to a source file.
- Allowed: The agent adds a fraud detector for card transactions.

## Latest change

0.0.4 · 2026-10-05 — Description now says what the policy does on Claude Code: warns. Behaviour unchanged.

## Provenance

- Tested by 4 of 11 cases run automatically
- SHA-256: `d40396513324363c19515b8d83f69114e88f7480188eb90d8842cc0e780d6303`
- Catalog commit: https://github.com/open-coder-ai/chock-catalog/commit/1e60d211c1bb767152c6ab0167932da06ab3a10f
- Source: https://github.com/open-coder-ai/chock-catalog/tree/1e60d211c1bb767152c6ab0167932da06ab3a10f/compliance/eu-ai-act-high-risk-triage (catalog path `compliance/eu-ai-act-high-risk-triage`)
- Presets: EU AI Act
- JSON: `/api/coding/policies/eu-ai-act-high-risk-triage.json`
- Report a problem: https://chock.sh/report/?policy=eu-ai-act-high-risk-triage&version=0.0.4&page=%2Fcoding%2Fpolicies%2Feu-ai-act-high-risk-triage%2F