capcut-editor

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill's shared utility file scripts/_common.py executes the shell command pgrep via subprocess.run. This is a legitimate safety check used to ensure the CapCut application is not running before the skill attempts to modify project files, which prevents data corruption.
  • [DATA_EXPOSURE]: The scripts/create-project.py script reads machine-specific identifiers (including device ID, hard disk ID, and MAC address) from existing local CapCut project files. This information is used to populate metadata for new projects to ensure they are compatible with the local CapCut installation. The data remains local and is not transmitted externally.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from CapCut project files, including project names and text overlay content. In scripts/edit-clips.py, this content is output to the terminal during timeline inspections. This represents a potential indirect prompt injection surface if a user-loaded project contains malicious instructions, though it is standard behavior for a project management tool.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 11:18 AM
Security Audit — agent-trust-hub — capcut-editor