record-episode
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the local script harness/scripts/mem-rank.py via the shell to process and store episodic memory entries.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied strings for session descriptions ('did'), outcomes, and session names, which are then interpolated into a shell command. This creates a potential surface for command injection if input contains shell metacharacters. * Ingestion points: User-provided values for 'did', '--outcome', and '--session' in SKILL.md. * Boundary markers: None defined to isolate user inputs within the command line string. * Capability inventory: Shell execution of python3 with variable arguments. * Sanitization: Instructions do not specify any validation or escaping for the user-provided inputs.
Audit Metadata