FFmpeg Video Editor
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed specifically to generate complex shell commands for the
ffmpegCLI. If the agent executing these commands does not properly sanitize filenames (e.g.,"; rm -rf / ; "), it could lead to arbitrary command execution on the host system. The use of-safe 0in the concatenation example specifically disables filename safety checks within FFmpeg, which is a known risk when processing untrusted file lists. - [INDIRECT_PROMPT_INJECTION]: The skill processes natural language requests to populate command templates. This creates an attack surface where a user (or data processed by the agent) could attempt to inject shell metacharacters or unintended flags into the generated command string.
- Ingestion points: Natural language user requests (SKILL.md)
- Boundary markers: None present in the instructions to the agent.
- Capability inventory: The skill generates commands for the
ffmpegtool which has full file system access and potential network capabilities. - Sanitization: The instructions rely on the LLM's inherent ability to extract parameters but do not provide explicit sanitization or escaping rules for shell-sensitive characters.
Audit Metadata