ai-ugc-test-loop
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection through the ingestion of external data.
- Ingestion points: The skill requests product URLs and between 3 to 10 external references, which can include links, transcripts, and captions (SKILL.md, Sections 1 and 2).
- Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested reference data.
- Capability inventory: The skill utilizes several media generation and account management tools, including
mcp__agent-media__make_ugc,mcp__agent-media__generate_video, andmcp__agent-media__get_account(SKILL.md frontmatter). - Sanitization: No explicit sanitization, filtering, or escaping of the external content is defined before it is processed by the agent.
- [SAFE]: All referenced tools are scoped to the vendor's namespace (prefixed with
mcp__agent-media__) and are consistent with the skill's stated purpose of media generation and account management. The instructions do not contain any evidence of malicious patterns, obfuscation, or persistence mechanisms.
Audit Metadata