The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and process SKILL.md files from sibling directories to answer user questions. This creates an attack surface where a malicious skill file could inject instructions into the agent's context.
Ingestion points: Reads SKILL.md files of other skills in adjacent directories.
Boundary markers: None specified in the instructions.
Capability inventory: Shell command execution via npx, git, gh, and file system operations (rm).
Sanitization: No sanitization or validation of the content from those files is mentioned.
[COMMAND_EXECUTION]: The skill provides and executes shell commands for managing skill life cycles, including npx skills, git, and gh. While the instructions include warnings against dangerous flags like --all that could lead to unintended file deletion, the agent is empowered to run these commands upon user request.
[EXTERNAL_DOWNLOADS]: The skill facilitates downloading code and configurations from the author's repository on GitHub and packages from the npm registry via npx.

wkc-help