erne-retrospective
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script at
.claude/hooks/scripts/evaluate-session.jsusing Node.js. - Evidence: The command
node .claude/hooks/scripts/evaluate-session.jsis invoked as part of the session analysis process. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from the current workspace.
- Ingestion points: The analysis script reads files created or modified during the session, test results, and build success/failure history.
- Boundary markers: There are no explicit boundary markers or instructions defined in the skill to treat the script's input or output as untrusted content.
- Capability inventory: The skill triggers a Node.js script execution and instructs the agent to "Act on recommendations" provided by that script.
- Sanitization: No evidence of input validation or sanitization is provided for the data consumed by the evaluation script.
Audit Metadata