architectural-amendment-protocol
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [NO_CODE] (INFO): The skill is composed entirely of markdown instructions without any associated scripts (Python, JavaScript, Bash) or executable binaries. Analysis is limited to the provided natural language prompts.
- [PROMPT_INJECTION] (SAFE): No patterns of instruction override, safety bypass, or role-play injection were found in the skill's instructions or metadata.
- [DATA_EXFILTRATION] (SAFE): There are no indicators of network requests, credential access, or sensitive file system operations.
- [EXTERNAL_DOWNLOADS] (SAFE): No external URLs or package dependencies were identified in the skill definition.
Audit Metadata