audience-research
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from YouTube comments.
- Ingestion points: The agent is instructed to 'mine comments' and 'read the last 50 comments' as part of Research Method 1 and Method 3 in SKILL.md.
- Boundary markers: The skill does not provide instructions to use delimiters or ignore embedded commands within the comments.
- Capability inventory: No shell commands, file-write operations, or network execution patterns are present in the skill files.
- Sanitization: The instructions do not specify any sanitization or filtering of the content retrieved from external comments.- [NO_CODE]: The skill consists entirely of natural language instructions in markdown format and does not contain any scripts, binaries, or shell command execution strings.
Audit Metadata