p-video-animate
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external URLs and prompt strings, creating an attack surface where malicious data could influence agent behavior. 1. Ingestion points: Found in SKILL.md, specifically the video, image, and instruction_prompt fields used in Predictions API calls. 2. Boundary markers: None identified; inputs are passed directly to API parameters. 3. Capability inventory: The skill utilizes network operations (curl) to communicate with api.pruna.ai. 4. Sanitization: No logic is provided to sanitize or validate the content of the provided URLs or prompts.
- [PROMPT_INJECTION]: The skill documentation includes an optional parameter that explicitly allows for bypassing safety controls. The Common optional fields section in SKILL.md lists disable_safety_checker as an available input for the generative process.
- [EXTERNAL_DOWNLOADS]: The skill references external resources for additional functionality. Instructions include installing various associated skills from the PrunaAI/pruna-skills repository on GitHub.
- [COMMAND_EXECUTION]: The skill provides instructions for the agent to use CLI tools to perform network requests. Example curl commands are documented for uploading source assets to api.pruna.ai/v1/files and creating predictions via api.pruna.ai/v1/predictions.
Audit Metadata