lab-notes
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by reading and processing user-controlled experiment log files to restore context and provide status reports.
- Ingestion points: The agent reads data from
experiments/*/LOG-*.mdfiles (specifically in the/lab-notesdispatcher and context restoration steps). - Boundary markers: There are no explicit delimiters or safety instructions (e.g., "ignore embedded commands") used when reading the log content back into the session.
- Capability inventory: The skill's workflow involves file creation, log appending, and directory management via
git mv. - Sanitization: The ingested log content is not sanitized or validated before being processed by the agent.
- [SAFE]: The skill incorporates laboratory record-keeping guidelines from trusted sources, including the National Cancer Institute (NCI) and Rice University, for user reference. These references are documented neutrally and point to established organizations.
Audit Metadata