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
Risk Level
LOW
Analyzed
Feb 16, 2026, 12:58 PM