The Agent Skills Directory

Indirect Prompt Injection (MEDIUM): The skill processes untrusted data (source code and comments) to generate documentation and has the capability to write to the filesystem.
Ingestion points: The agent uses Read, Grep, and Glob tools to ingest data from the workspace.
Boundary markers: No delimiters or safety instructions are provided to help the agent distinguish between data to be documented and potential malicious instructions embedded in the source files.
Capability inventory: The agent is granted Write and Edit tools, allowing it to modify the file system based on potentially poisoned input.
Sanitization: No sanitization or validation mechanisms are defined for the input content.
Command Execution (LOW): The inclusion of Write and Edit tools allows for direct modification of files. While the description specifies the docs/ folder, the tool access is not technically restricted within the skill's permission metadata.

docs-api