The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill defines a command to run a local script: python scripts/api_validator.py <project_path>. This involves running Python code against user-provided file paths.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill's primary function (validating project APIs) requires reading untrusted data from an external project path. This creates a surface for indirect prompt injection. Mandatory Evidence Chain: 1. Ingestion points: Files within the user-provided <project_path> are read and processed. 2. Boundary markers: Absent; the skill does not specify delimiters or instructions to ignore embedded prompts in the analyzed project. 3. Capability inventory: The skill has Read, Write, Edit, Glob, and Grep permissions, and executes a Python subprocess. 4. Sanitization: No evidence of input sanitization or validation of the content read from the project path is present in the skill metadata.

api-patterns