resume
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Bash tool to execute local system commands to retrieve session metadata. Evidence: Execution of 'date', 'cat', and 'ls' to display context.
- [PROMPT_INJECTION]: The skill processes untrusted data from local session files which could contain malicious instructions. Ingestion points: Reads files from '.claude/sessions/'. Boundary markers: None. Capability inventory: Bash, Read, Write tools. Sanitization: None.
Audit Metadata