Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/lexbeam-software/eu-ai-governance-pluginWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/lexbeam-software/eu-ai-governance-plugin/classify-ai-risk)<a href="https://agentmods.dev/commands/lexbeam-software/eu-ai-governance-plugin/classify-ai-risk"><img src="https://agentmods.dev/badge/commands/lexbeam-software/eu-ai-governance-plugin/classify-ai-risk.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00023 | $0.00520 |
| Opus 5 | $0.00012 | $0.00260 |
| Sonnet 5 | $0.00005 | $0.00104 |
| Haiku 4.5 | $0.00002 | $0.00052 |
Grade A, and why
classify-ai-risk scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 33 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/eu-ai-governance:classify-ai-risk
Classify the supplied system or GPAI model using the ai-act-classification skill. Follow the legal source protocol. See connected sources.
Workflow
- Ingest evidence. Read supplied descriptions and documents. Identify date, version, author, and whether each statement is fact, vendor claim, or assumption.
- Ask only decisive questions. Establish intended purpose, affected persons, decision effect, roles, geography, data, profiling, biometrics, synthetic content, workplace use, Annex I product-law context, and exact Annex III use.
- Ground the analysis. When available, call
euaiact_classify_system; useeuaiact_get_articlefor a versioned summary and official URL, then read the complete official provision; useeuaiact_assess_art6_3_exceptionfor a possible Annex III exception; calleuaiact_get_obligationsfor provider and deployer roles only, map other actor duties from the official text, and calleuaiact_check_deadlines. - Apply every gate. Scope and role; all Article 5 points; Article 6(1) including 1a to 1c; Article 6(2) and exact Annex III item; Article 6(3); Article 50; GPAI Articles 51 to 55; national and sector issues.
- Challenge the answer. Identify the strongest alternative classification and the fact or legal boundary that resolves it. Read the complete provision before making a negative legal claim.
Output
Return:
- executive classification, allowing parallel labels
- scope, facts, assumptions, and missing decisive evidence
- gate-by-gate table with exact citation and rationale
- Article 6(3) assessment where relevant
- obligations by actor and operative date
- evidence requests and no more than five next actions
- confidence and source note
Use not established on supplied facts when evidence is incomplete. Do not claim certification, guaranteed compliance, or regulator acceptance.
With --lang de, write idiomatic German with umlauts and preserve official EU-law terms and citations.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 33 lines · 23 tokens per session scan A 74ea6b441b7e
classify-ai-risk is a command published in the GitHub repository lexbeam-software/eu-ai-governance-plugin (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 23 tokens to every session and 520 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
tcop
Generate a Technology Code of Practice (TCoP) review document for a UK Government technology project.
evidence
Export a signed compliance evidence package as JSON or PDF for regulatory handoff.
comply
Run a full EU AI Act compliance scan on your Python AI project.
ai-act-scan
Scan a codebase for EU AI Act compliance evidence and gaps. Produces a dimension-scored report with per-file findings, architecture graph, and prioritized recommendations.
ai-act-article
Show which analyzers, compliance dimensions, and current findings in this codebase map to a specific EU AI Act article.
ai-act-scan-fix
Scan a codebase, then propose concrete remediation (code edits, new files, tests) for the top compliance gaps. Does NOT auto-apply — always shows the plan first.