The Agent Skills Directory

[SAFE]: The skill performs legitimate workspace management and template generation tasks. It includes robust safety features such as a mandatory collision plan and explicit user approval requirements before modifying or overwriting any existing files. The skill also implements privacy-conscious practices, such as using SHA-256 hashes of local paths for tracking markers instead of storing the raw paths.
[COMMAND_EXECUTION]: The skill instructs the agent to use local tools (such as icm validate) to verify the integrity of the generated workspace. These commands are executed within the context of the skill's primary purpose and do not involve untrusted remote sources or arbitrary code execution.

icm-grill