agent-pseudocode
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill uses local shell commands in lifecycle hooks for logging and timestamping.
- Evidence: Hooks in
SKILL.mduseecho,date, andtail. - Context: These commands are hardcoded and used for phase management, posing no risk of unauthorized execution or privilege escalation.
- [DATA_EXFILTRATION] (SAFE): No patterns of data exfiltration, network requests, or hardcoded credentials were found.
- [PROMPT_INJECTION] (SAFE): No instructions were detected that attempt to override agent safety protocols or bypass constraints.
- [EXTERNAL_DOWNLOADS] (SAFE): The skill does not define external dependencies or download remote code.
Audit Metadata