cold-open-creator
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process podcast transcripts, which are external and potentially untrusted data sources. This creates a surface for indirect prompt injection where instructions hidden within a transcript could attempt to influence the agent's logic or output.
- Ingestion points: The skill instructions in
SKILL.mddirect the agent to scan and extract clips from a provided transcript (Step 1 and Step 2). - Boundary markers: The instructions do not define clear delimiters or include a directive for the agent to ignore any natural language instructions found within the transcript text.
- Capability inventory: The skill's primary capabilities are limited to text selection, cutting, and rearranging, which mitigates the risk of high-severity actions (like file deletion or network requests) resulting from an injection.
- Sanitization: There is no evidence of sanitization, filtering, or validation of the input transcript to prevent instructions from being processed as part of the content selection logic.
Audit Metadata