The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads external source code and design documents to identify logical sections and generate explanatory comments. Because it processes untrusted input and has the ability to write to the filesystem, it is susceptible to indirect prompt injection where malicious instructions hidden in the files could influence the agent's output or actions.
Ingestion points: The skill reads user-provided files via <file-path> and an optional --context <path> using the Read tool.
Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings for the agent when reading the target files.
Capability inventory: The skill uses Read, Edit, Write, Glob, and Grep tools, which allows for potentially widespread file modification if the agent is manipulated.
Sanitization: There is no mention of sanitizing, escaping, or validating the content of the ingested files before the agent processes them.

literatize