The Agent Skills Directory

[Indirect Prompt Injection] (SAFE): The skill processes untrusted data from output/TUTORIAL_SPEC.md (Ingestion point). No boundary markers are present (Boundary markers). Its capabilities are limited to reading and writing text files (Capability inventory). No content sanitization is specified (Sanitization). However, the absence of network access and shell execution prevents any meaningful exploitation of this ingestion surface.
[Data Exposure & Exfiltration] (SAFE): No access to sensitive system paths or credentials was identified. The skill only interacts with the local tutorial directory.
[Remote Code Execution] (SAFE): This skill contains no executable code or calls to external scripts. It functions entirely through natural language instructions for the AI agent.

concept-graph