The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and modify external data (code comments) and possesses file-writing capabilities, which creates a potential surface for indirect prompt injection.
Ingestion points: The skill processes code files during editing or writing tasks, as specified in the 'How to use this skill' section.
Boundary markers: The instructions include a 'Scope guardrail' and a 'Code Comment Checklist' to limit the agent's actions to comment modification only.
Capability inventory: The skill is permitted to use Read, Write, Edit, Glob, and Grep tools, allowing it to inspect and modify files.
Sanitization: There is no explicit sanitization of the input code content, although the skill's primary function is to filter out specific patterns of text.

antislop-code