The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses find and ls commands to identify local skill configuration files (SKILL.md) within the .claude/skills and plugins directories. These are read-only discovery operations used to help the agent reference existing tools.
[COMMAND_EXECUTION]: Uses the ln -s command to create a symbolic link between AGENTS.md and CLAUDE.md, which is a standard practice for maintaining project-level documentation.
[PROMPT_INJECTION]: The skill processes untrusted local data by reading found SKILL.md files to generate documentation. While this represents an indirect prompt injection surface where a malicious local file could influence the agent's summary, the risk is considered minimal as it is limited to the local file system and the intended documentation task.

agents-md