self-media-short-video
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content (scripts, video plans) and outputs from third-party tools (AI voice/video synthesis) to generate complex production files and publishing metadata. The lack of explicit boundary markers or sanitization logic creates a surface where embedded instructions in the source text could influence the agent's behavior during the multi-step production process.
- Ingestion points: User-supplied scripts, platform-specific metadata, and outputs from external video/audio synthesis tools in the video production pipeline (SKILL.md, animated-explainer-production.md).
- Boundary markers: Absent; the instructions do not specify delimiters or "ignore embedded instructions" warnings for untrusted input.
- Capability inventory: The skill is designed to coordinate with external tools (Topview MCP), write production files to the file system, and pass data to publishing agents (digital-human-production.md).
- Sanitization: Absent; the skill relies on the agent's general reasoning to handle unsafe content without specific filtering protocols.
Audit Metadata