continuous-learning

Warn

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a shell script (evaluate-session.sh) to perform session analysis and directory management.
  • [COMMAND_EXECUTION]: The documentation instructs users to configure a persistent 'Stop' hook in ~/.claude/settings.json. This ensures the evaluator script executes automatically after every session, establishing a permanent persistence mechanism for local code execution.
  • [DATA_EXFILTRATION]: The evaluator script reads session transcripts provided via the transcript_path. This grants the skill access to the complete history of user interactions and agent responses, which constitutes sensitive data exposure.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 19, 2026, 06:51 AM
Security Audit — agent-trust-hub — continuous-learning