Add AI Act Companion (open-source classifier) under Risk Assessment & Classification

[JKasteele]

Thanks for maintaining this list!

Adding AI Act Companion under Open-Source Projects → Risk Assessment & Classification — a free, MIT-licensed, local-first EU AI Act risk classifier.

Why it fits the open-source section:

• Deterministic & cited — every tier verdict links to the exact Article/Annex; a rule engine, not an LLM.
• Architecture-aware AI-security lens — OWASP LLM Top 10 + MITRE ATLAS + STRIDE, severity driven by the system architecture, plus a red-team plan and matching controls.
• Also generates DPIA, Annex IV technical docs, FRIA, a conformity tracker (Art. 99 penalties) and a post-market monitoring plan; NIST AI RMF + ISO/IEC 42001 crosswalks.
• Live demo (synthetic data only); tests + CI; also ships as a Claude Code plugin and a GitHub Action.

Entry added at the end of the subsection, matching the existing format. Happy to tweak wording/placement — thanks for considering!

[carloshvp]

Thanks for the contribution. The project looks relevant and the demo link works, but the repository is very new and I could not find visible independent adoption yet.

Before inclusion, could you provide maturity or usefulness evidence such as external users, package/download usage, independent contributors, citations, integrations, deployments, or examples validated outside the project team?

The proposed entry is also too long and claim-heavy for the list. Please shorten it to one neutral sentence focused on the core EU AI Act use case. For example:

- [AI Act Companion](https://github.com/JKasteele/ai-act-companion) - Local-first risk classifier for EU AI Act obligations, with cited outputs and generated compliance documents.

A concise entry plus credible maturity evidence is needed before merge.

[carloshvp]

Final reminder: I have not received a response to the review from 27 June. The PR still proposes a 362-character entry and does not provide evidence of use outside the project team.

Please replace the entry with one neutral sentence under 100 characters and reply with maturity or usefulness evidence by 4 August 2026. If there is no response by then, I will close the PR to keep the review queue clean. You may open a new PR later once the wording and evidence are ready.