The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external project files through tools like verse_read_file (SKILL.md), which creates an attack surface for instructions embedded in source code. This is an inherent part of its development purpose and is handled with appropriate procedural warnings.
Ingestion points: verse_read_file is used to inspect project files (SKILL.md, EXAMPLES.md).
Boundary markers: None specified for the read content.
Capability inventory: Includes verse_write_file, verse_create_device, and execute_python (SKILL.md).
Sanitization: None specified.

verse-uefn