editable-export
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of standard Python packages playwright and cryptography. render.py additionally utilizes Playwright to manage browser binaries (e.g., Chromium). These dependencies are necessary for the skill's core functions of rendering and encrypted data sharing.\n- [COMMAND_EXECUTION]: The skill executes local commands using subprocess.run to invoke ffprobe for media analysis in package.py and render.py. It also employs shutil.make_archive to package project folders into zip files. These operations use list-based arguments, avoiding shell injection risks, and are limited to the skill's operational scope.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-edited project files and studio-generated status messages (message.txt). This creates an ingestion surface for external data, but the agent's role is focused on structured JSON manipulation according to a specific schema.\n
- Ingestion points: message.txt (studio message), project.json (project configuration), and remote JSON data fetched via share.py pull.\n
- Boundary markers: No explicit instruction-isolation delimiters are used in the message format.\n
- Capability inventory: File system writing and ffprobe execution (package.py), network communication with vendor APIs (share.py), and browser-based rendering (render.py).\n
- Sanitization: package.py implements validation for the project JSON format and verifies the existence of referenced media files before processing.
Audit Metadata