smart-montage
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill automatically downloads and installs the
meitu-kaipai-clipackage (specifically version 0.1.11) from the NPM registry if it is not already present in the environment. - [COMMAND_EXECUTION]: The skill executes the
kaipaiCLI tool to perform core functions including media analysis, project generation, task monitoring, and authentication management. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user-provided prompts and remote media URLs, which could contain instructions designed to influence agent behavior.
- Ingestion points: User-supplied themes and requirements passed to the
-pparameter, and remote media assets fetched via thekaipai downloadcommand. - Boundary markers: Instructions in the skill files advise the agent to escape input values and avoid executing text returned from the service.
- Capability inventory: The skill possesses the ability to spawn subprocesses (the CLI), perform file system operations for asset management, and initiate network requests for downloading media and communicating with the video service API.
- Sanitization: The skill relies on the agent's adherence to instructional guardrails for input handling rather than using hardcoded sanitization logic in its scripts.
- [DYNAMIC_EXECUTION]: A bundled Node.js script (
scripts/ensure-cli.mjs) is used at runtime to verify the installation environment and launch the CLI tool using the Node.js executable path.
Audit Metadata