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/Peaky8linders/eu-ai-act-scannerWrote 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/peaky8linders/eu-ai-act-scanner/ai-act-ask)<a href="https://agentmods.dev/commands/peaky8linders/eu-ai-act-scanner/ai-act-ask"><img src="https://agentmods.dev/badge/commands/peaky8linders/eu-ai-act-scanner/ai-act-ask/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/peaky8linders/eu-ai-act-scanner/ai-act-ask"><img src="https://agentmods.dev/badge/commands/peaky8linders/eu-ai-act-scanner/ai-act-ask.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00046 | $0.00548 |
| Opus 5 | $0.00023 | $0.00274 |
| Sonnet 5 | $0.00009 | $0.00110 |
| Haiku 4.5 | $0.00005 | $0.00055 |
Grade A, and why
ai-act-ask 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 12d 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.
What it actually says
/ai-act-ask
Answer a compliance question, grounded in the bundled EU AI Act knowledge base: verbatim article/annex text (Regulation (EU) 2024/1689), concise obligation paraphrases, and the four-axis compound-risk taxonomy. This is the plugin's local "Lexy" retrieval-and-answer path.
Arguments
$1(required) — the question, e.g.What does Article 50 require for deepfakes?--top-k N(default 4) — number of grounded sources to retrieve.
Behaviour
- Run the grounded retriever:
It prints the answer, the cited articles, the related compliance dimensions, and the grounded source excerpts. It runs 100% locally — no LLM, no network — by default.python -m scanner.cli --ask "$1" - Present the answer and ALWAYS keep the citations. Do not add obligations that the cited articles do not support.
- In
assistedmode (checkpython -m scanner.cli --settings), synthesise a fuller narrative answer yourself over the returned sources — you are the user's Claude Code — but treat the retrieved statute text as ground truth and cite every claim by article. For headless CLI use,--mode assistedroutes synthesis through the Claude Max wrapper and applies a citation guard. - For code-specific questions ("does MY repo satisfy Art. 50?"), run
/ai-act-scanfirst and combine its findings with the grounded answer.
Example
User: /ai-act-ask What are a deployer's Article 50 transparency duties?
Response cites Art. 50(3) (emotion/biometric notice) and Art. 50(4) (deep-fake +
public-interest-text disclosure), notes these are deployer obligations
enforceable 2 August 2026, and links them to the content_transparency
dimension — grounded verbatim in the bundled statute text.
Notes
- Grounding corpus: Regulation (EU) 2024/1689 verbatim text (EUR-Lex CELEX 32024R1689) + obligation paraphrases + the agentic-risk taxonomy.
- This is compliance information, not legal advice.
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.
- 12d ago First seen · 51 lines · 46 tokens per session scan A 7e67472e62af
ai-act-ask is a command published in the GitHub repository Peaky8linders/eu-ai-act-scanner (5 stars, last pushed 5d ago), licensed Apache-2.0. It adds 46 tokens to every session and 548 once invoked, about $0.0002 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
discover
Discover runtime AI components and static AI dependencies with AIR Blackbox.
evidence
Export a signed compliance evidence package as JSON or PDF for regulatory handoff.
replay
Replay recorded AI traces to detect behavioral drift between model versions.
comply
Run a full EU AI Act compliance scan on your Python AI project.
status
Check the current governance status for your session, including your plan, remaining capacity, and whether the governance layer is operating normally.
audit
Generate a structured audit report for the current session or a specified time range. The report summarises every governed action, lists any blocked requests, and provides a compliance-ready summary you can download or share.