mcaf
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references documentation and tutorial resources at
https://mcaf.managed-code.com/and the source repository athttps://github.com/managedcode/MCAF. These are recognized official resources for the managedcode organization. - [COMMAND_EXECUTION]: The skill instructs the agent to run
scripts/verify-mandatory-policies.shto validate that the repository adheres to the MCAF policy contract. This is a local repository-native validation step. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests untrusted configuration data from
AGENTS.mdfiles within the repository. Ingestion points: Reads root and localAGENTS.mdfiles. Boundary markers: Instructions require reading files completely and merging new guidance without weakening or omitting existing rules. Capability inventory: Executes local validation scripts and writes to governance-related markdown files. Sanitization: No explicit sanitization or escaping of ingested markdown content is defined beyond the logical merging instructions.
Audit Metadata