The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions and examples for using local shell commands like rg (ripgrep) and find to inventory files and verify directory integrity. These commands are benign and used for local discovery purposes within the development environment.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a development tool that processes other skill files, creating a potential surface for indirect prompt injection from external or user-provided content. 1. Ingestion points: Contents of SKILL.md, references/, and examples/ directories of target skills being created or repaired. 2. Boundary markers: No explicit delimiters or boundary instructions are mentioned to isolate the content of the skills being processed from the agent's instructions. 3. Capability inventory: The skill has the capability to read and write files in the local workspace and execute environment discovery tools. 4. Sanitization: No explicit validation, filtering, or sanitization protocols are defined for handling the content of skills being managed.

skill-development