session-search
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: Instructs the agent to use the system
greputility to perform text searches across local JSONL files located in~/.feynman/sessions/. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and process untrusted data from historical session transcripts.
- Ingestion points: Processes files from the
~/.feynman/sessions/directory. - Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the transcript search workflow.
- Capability inventory: Includes shell execution capabilities (bash) used for searching.
- Sanitization: No evidence of sanitization or validation of the historical transcript content prior to processing.
Audit Metadata