giggle-generation-video
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data provided by the user via command-line arguments to generate videos. \n
- Ingestion points: User-defined video prompts and reference frames (URLs or Base64 data) passed via
--prompt,--start-frame, and--end-frameinSKILL.mdandscripts/generation_api.py.\n - Boundary markers: None present to distinguish instructions from data in the interpolated prompts.\n
- Capability inventory: Network communication with the vendor API and local script execution using
python3.\n - Sanitization: The script includes a basic check for the 'sensitive' keyword in API responses to handle content filtering results from the server.\n- [COMMAND_EXECUTION]: The skill uses the agent's capability to execute a local Python script (
scripts/generation_api.py) to manage task submission and status polling.\n- [SAFE]: The skill interacts exclusively with the vendor's official domaingiggle.proand uses standard environment variables for secret management, adhering to best practices for API-based skills.
Audit Metadata