create-control-manifest
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs legitimate documentation tasks by reading local project files (ADRs, technical preferences) and writing a consolidated manifest to the architectural documentation directory.
- [SAFE]: Strong human-in-the-loop safeguards are implemented via
AskUserQuestion, requiring user approval for the rule summary, the technical director's review findings, and the final file write operation. - [SAFE]: The skill uses a multi-agent workflow, spawning a
technical-directoragent to perform a quality and safety review of the extracted rules, which serves as a validation gate. - [DATA_EXPOSURE]: The skill processes external, user-contributed content from the repository (ADRs and technical preferences) as its primary input. While this creates a surface for indirect prompt injection, the risk is minimized by the specific extraction requirements, the secondary agent review process, and final human verification.
Audit Metadata