yt-comment
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process YouTube comment threads, which are untrusted external inputs. A malicious comment could contain instructions intended to override the agent's behavior, such as forcing it to generate inappropriate replies or alter the creator's voice profile.
- Ingestion points: Untrusted data enters the context via "pasted comment threads" or the agent's own data fetching when instructed to "handle the comment section."
- Boundary markers: The skill lacks explicit instructions for the agent to treat the ingested comment text as strictly data, such as using delimiters or 'ignore embedded instructions' warnings.
- Capability inventory: The skill reads from and writes to a local configuration file at
~/.codex/youtube/voice.mdand generates text drafts for the user. - Sanitization: There is no mention of sanitization, filtering, or validation of the external comment content before it is processed to infer voice or draft replies.
Audit Metadata