add-camera-controls
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill is designed to fetch code and configuration from an external MCP server named 'threenix'. It executes this by writing the returned file content directly to the local file system and integrating it into the application logic.
- [COMMAND_EXECUTION]: The skill performs command-line operations including package installation using the project's existing package manager and running build or type-check commands to verify the integration.
- [DYNAMIC_EXECUTION]: The skill dynamically generates project files by decoding and writing content (including Base64 data) provided by the external tool at runtime.
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it parses external repository instructions and configuration data.
- Ingestion points: Reads local repository instructions and package manifests (SKILL.md, Step 1).
- Boundary markers: None defined to isolate the processed data from instructions.
- Capability inventory: Includes file writing, package installation, and execution of build/type-check commands (SKILL.md, Steps 4, 6, 7).
- Sanitization: The instructions do not specify validation or sanitization of the data read from the repository.
Audit Metadata