muapi-social-media-video
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from local workspace files (brand-identity.md, ICP.md, messaging.md) to guide the generation of social media content and technical video prompts.
- Ingestion points: SKILL.md (Step 1) specifies reading three brand files from the working directory or brand/ subdirectory.
- Boundary markers: The instructions do not define delimiters or provide warnings to the agent regarding potential instructions embedded within these brand files.
- Capability inventory: The skill invokes local shell scripts (generate-image.sh and run-social-video.sh) which can execute complex logic.
- Sanitization: There is no explicit sanitization of the content read from these files before it is used to construct prompts.
- [COMMAND_EXECUTION]: The skill instructs the agent to construct and execute shell commands where significant parameters (like the --prompt argument) are dynamically generated by the AI based on the processed brand data.
- Evidence: Step 4 and Step 5 in SKILL.md provide templates for executing bash scripts where the AI-generated 'director brief' or 'image prompt' is passed directly as a command-line argument.
Audit Metadata