The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of analyzing and rewriting instructions from external sources.\n
Ingestion points: Data is ingested from files via the Read tool or from user-provided text in the conversation context, as specified in the <workflow> and <input_handling> sections.\n
Boundary markers: The instructions mention preserving placeholders and delimiters but do not include explicit guardrails or "ignore" instructions for the content being processed.\n
Capability inventory: The skill uses Read, Write, Edit, Glob, and Grep tools, allowing it to interact with the file system based on potentially poisoned input.\n
Sanitization: There is no mention of sanitizing, escaping, or validating the input content to prevent the agent from executing instructions found within the data.

prompt-doctor