The Agent Skills Directory

[SAFE]: The skill's primary purpose is to establish security guardrails by allowing the agent to create rules that monitor for dangerous patterns in commands and files. It provides examples of sensitive files (.env, credentials) and dangerous commands (sudo, rm -rf) as patterns for the agent to watch and warn against.
[INDIRECT_PROMPT_INJECTION]: The skill manages a configuration system where rule files are processed to influence agent behavior. While this creates a mechanism for behavioral modification through warnings and blocks, it is explicitly designed for user-authorized guardrails and does not exhibit malicious intent.

hookify-rules