talk-stage4-position
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external markdown files which could theoretically contain malicious instructions targeting the agent's logic.
- Ingestion points: The skill reads data from
talks/{YYYY}-{slug}-summary.mdandtalks/{YYYY}-{slug}-concepts.md. - Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between the agent's instructions and the content of the talk files.
- Capability inventory: The skill uses the
Writetool to create local files andAskUserQuestionto interact with the user. It does not have access to the network or shell execution. - Sanitization: No evidence of input validation or escaping for the ingested content was found in the skill logic. However, the risk is mitigated by a mandatory human-in-the-loop checkpoint before further stages.
Audit Metadata