MCP-first AI foundation for TYPO3. Bring AI agents into TYPO3, on your terms.
Machine-readable AI permissions for websites. A consolidated spec at /.well-known/ai-policy.json for declaring how AI agents may train on, search, or use your content.
A scenario library and testing framework for GenAI alignment — objective drift, robustness, and agentic/enterprise risk — tested against Azure OpenAI and Claude, with audit-ready findings mapped to governance frameworks.
**The data behind agent permission rules.** Verified facts about what each CLI command does — destructive flags, network reach, BSD vs GNU differences — with a re-runnable check behind every row, so a policy is written from evidence instead of intuition.
To associate your repository with the ai-permissions topic, visit your repo's landing page and select "manage topics."