compress-video

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill directs the agent to execute local shell commands such as find, ls, stat, and ffmpeg for identifying and re-encoding video files. This is consistent with its stated functionality.
  • [PROMPT_INJECTION]: The skill processes user-controlled file paths, creating a surface for indirect prompt injection. 1. Ingestion points: User-provided directory and file paths are used to construct shell commands. 2. Boundary markers: No explicit delimiters are used for filename inputs. 3. Capability inventory: The skill has access to shell execution via subprocess calls. 4. Sanitization: The instructions mitigate risks by explicitly requiring the agent to quote variables representing user input.
  • [SAFE]: No malicious patterns, such as data exfiltration or persistence mechanisms, were detected. The skill uses standard system tools for their intended purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 11:22 AM
Security Audit — agent-trust-hub — compress-video