claude-video-generate

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute local scripts, specifically scripts/video_generate.py and scripts/preflight.sh, to interface with video generation APIs and validate the environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions create a potential vulnerability surface by interpolating user-supplied text directly into shell commands via the --prompt argument.
  • Ingestion points: User-provided video descriptions in SKILL.md are passed as arguments to shell-executed Python scripts.
  • Boundary markers: While prompts are enclosed in double quotes in the examples, there are no instructions to escape or sanitize shell-sensitive characters like backticks (`), semicolons (;), or command substitution syntax ($()).
  • Capability inventory: The agent has access to the Bash tool and file system (Read, Write).
  • Sanitization: The instructions do not define any validation, filtering, or escaping protocols for external user input before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 06:38 PM
Security Audit — agent-trust-hub — claude-video-generate