The Agent Skills Directory

[SAFE]: The skill is a development tool focused on critiquing and improving project documentation. Its operations are limited to reading local files and interacting with the user.
[PROMPT_INJECTION]: The skill processes untrusted content from local files (plans, specs, and RFCs), creating an attack surface for indirect prompt injection. While a malicious file could theoretically contain instructions to subvert the agent, the skill itself provides a benign utility with no built-in malicious logic.
Ingestion points: Accesses local files using the Read, Glob, and Grep tools based on the provided $ARGUMENTS in SKILL.md.
Boundary markers: The skill does not use specific delimiters or instructions to ignore embedded commands within the files it reads.
Capability inventory: The skill utilizes the Write, Edit, and AskUserQuestion tools to update plans and communicate with the user.
Sanitization: No content validation or sanitization is performed on the data ingested from local files.

interview