filmcraft-cli-audio

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads binary runtimes and desktop applications from GitHub repositories (storytold/filmcraft and full-aigc-skills/filmcraft-skills). These external resources are strictly controlled via lock files (runtime.lock.json, desktop.lock.json) that specify exact URLs, expected byte sizes, and SHA256 hashes to prevent tampering during delivery.
  • [COMMAND_EXECUTION]: The skill executes the downloaded CLI tools and various macOS system utilities (like hdiutil, codesign, and lipo) to manage installation and sessions. To mitigate the risks of arbitrary code execution, all subprocess calls are wrapped in a macOS sandbox profile (scripts/execution_permissions.py) that enforces least-privilege access, restricting the process to specific authorized project directories and denying unauthorized network or system resource access.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined task plans in JSON format (craft-command-plan/v1). It implements comprehensive validation logic in scripts/commands.py and scripts/workflow.py, including strict JSON key uniqueness checks, parameter type validation (e.g., 64-bit integer 'ticks'), and a whitelist of allowed command IDs derived from a local catalog to prevent the execution of unauthorized or malicious instructions embedded in project data.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 09:00 PM