The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to read and process markdown files from the user's workspace which could contain embedded instructions for the agent.\n
Ingestion points: Triggered by and processes .md files in docs/ and README.md as specified in SKILL.md.\n
Boundary markers: There are no instructions to use delimiters or to disregard instructions found within the input files.\n
Capability inventory: The agent has the ability to write and edit files on the local filesystem, which could be abused if an injection is successful.\n
Sanitization: No sanitization or filtering is performed on the content of the markdown files before processing.\n- [SAFE]: The provided Python script scripts/validate_markdown.py is a standard linting tool that checks heading hierarchy, code blocks, and link syntax. It uses only built-in libraries and performs safe file read operations without network access or execution of file content.

technical-writer