distill-sessions
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various system utilities including
find,jq,awk,perl,xargs,stat,sort, andheadto locate and extract data from local log files. - [INDIRECT_PROMPT_INJECTION]: The skill processes raw session transcripts which are considered untrusted external data that may contain malicious instructions designed to influence the agent's analysis.
- Ingestion points: Accesses JSONL log files located in
~/.claude/projects/and~/.codex/sessions/. - Boundary markers: Relies on
jqquery logic to isolate specific fields (user messages, tool commands, errors) but lacks explicit delimiters to prevent the agent from potentially following instructions found within the log content. - Capability inventory: Extensive use of shell processing tools and the ability to dispatch sub-agents to process batches of logs.
- Sanitization: The skill includes a mandatory redaction step to replace secrets like API keys, emails, and passwords with a placeholder before outputting the results.
Audit Metadata