Add CorvinOS — structural EU AI Act + GDPR enforcement for AI agents

[veegee82]

What

Adds CorvinOS to the AI Agent Governance section.

Why it fits

CorvinOS is the only entry in this list that enforces multiple EU AI Act and GDPR articles as structural runtime constraints rather than documentation or post-hoc compliance tools:

Article	Mechanism	Enforcement property
EU AI Act Art. 50	Bot-disclosure card	One-time per user · structurally fail-closed · no bypass path
EU AI Act Art. 5	Acceptable-use gate (L44)	SHA-256-anchored policy · no kill-flag · no tenant override
GDPR Art. 6/7	Consent gate	Deny-by-default · TTL-capped · re-validated at every turn
GDPR Art. 17	Erasure orchestrator	Cross-layer · pseudonymous subject IDs · audit trail de-linked
GDPR Art. 30/32	Hash-chained audit log	SHA-256 chain · daily auto-verify · chain write failure blocks request

None of these can be disabled by an environment variable or config flag — they are load-bearing code.

CorvinOS also bridges Ollama and cloud LLMs to Discord, Telegram, WhatsApp, Slack, and Email, making it a practical self-hostable deployment target, not just a reference implementation.

• GitHub: https://github.com/CorvinLabs/CorvinOS (Apache-2.0, public since 2026-06-28)
• PyPI: https://pypi.org/project/corvinos/ (pip install corvinos)

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[carloshvp]

Thanks for the contribution. The project now shows engineering maturity through package releases, tests and CI, 22 stars, and 4 forks. The proposed entry still exceeds the list format at 247 characters and makes legal and architecture claims that this list cannot verify.

Please replace it with a neutral description under 100 characters:

- [CorvinOS](https://github.com/CorvinLabs/CorvinOS) - Self-hosted agent platform with consent controls, audit logs, and AI disclosure features.

Please also provide one or two signs of use outside the project team, such as deployments, external users, integrations, or contributions. Release volume and stars are useful signals, but they do not establish independent usefulness on their own for a new self-submitted project.

The automated usage-limit comment above does not represent a maintainer review.