continuous-learning
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The shell script
evaluate-session.shexecutes standard system utilities includingmkdir,grep, andjqto manage local directories and verify session metadata.- [PROMPT_INJECTION]: The skill's primary function is to process session transcripts to extract 'learned skills', which creates a surface for indirect prompt injection. - Ingestion points:
evaluate-session.shreads session transcripts from the path provided in theCLAUDE_TRANSCRIPT_PATHenvironment variable. - Boundary markers: Absent; the skill does not use delimiters or explicit instructions to ignore potentially malicious content within the session transcript.
- Capability inventory: The skill can write to the local filesystem (
mkdir) and has read access to environment variables. - Sanitization: No sanitization or validation of the transcript content is performed before the agent is signaled to evaluate it for pattern extraction.
Audit Metadata