The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes markdown content from external files which may contain hidden instructions designed to influence the behavior of the AI agent.
Ingestion points: The markdown.js script reads the content of user-provided files using fs.readFileSync in the main function.
Boundary markers: The script does not utilize delimiters or specific instructions to help the agent distinguish between processed markdown data and the agent's own control instructions.
Capability inventory: The skill has file system access, specifically the ability to read files (fs.readFileSync) and write files (fs.writeFileSync) as demonstrated in markdown.js.
Sanitization: No sanitization or filtering is performed on the input content to remove or neutralize potential natural language instructions before they are processed by the agent.

markdown-tools