social-producer-agent
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes local Python scripts (e.g., gemini.py, veo.py) and the ffmpeg utility via the command line to generate and process media. These commands are constructed using variables and user-supplied parameters, which is a standard but sensitive operation.
- [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface where untrusted data from user responses is interpolated into shell commands.
- Ingestion points: User input gathered via the AskUserQuestion tool in SKILL.md (specifically platform names, quantities, themes, prompts, and asset paths).
- Boundary markers: None present in the skill instructions to delimit user input from the shell command structure.
- Capability inventory: Shell execution of multiple Python scripts (gemini.py, veo.py, lyria.py, video_audio_merge.py) and the ffmpeg binary as documented in SKILL.md.
- Sanitization: No explicit sanitization, shell-escaping, or validation of user input is mentioned before the data is interpolated into --prompt or --reference arguments, potentially allowing a malicious user to execute arbitrary commands via crafted input.
Audit Metadata