The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill is highly vulnerable to indirect prompt injection because it is designed to ingest and analyze untrusted content from a user's project directory while possessing powerful write and execution tools.
Ingestion points: Project files and directory structures accessed via the Read, Glob, and Grep tools, as well as the input processed by api_validator.py.
Boundary markers: Absent. The skill provides no instructions to the agent to treat external file content as data rather than instructions.
Capability inventory: The skill uses Write and Edit tools and executes a Python script, providing a path for malicious instructions in a project file to trigger unauthorized code changes or system commands.
Sanitization: Absent. There is no evidence of input validation or content sanitization before processing project data.
[Command Execution] (MEDIUM): The skill defines a command to execute a local Python script api_validator.py with a user-supplied <project_path>. If the script does not properly sanitize this input, it could be vulnerable to command injection or path traversal.

api-patterns