The Agent Skills Directory

[SAFE]: The skill implements a safety mechanism to prevent agents from performing dangerous actions without user consent. It defines a protocol for classifying decisions based on a local registry or pattern matching.
[COMMAND_EXECUTION]: The skill contains grep commands in the Procedure section (grep -niF ..., grep -niE ...). These are used for internal data lookup within the skill's own reference files (decision-registry.md, patterns.md) and do not pose a risk as they operate on static, bundled content.
[NO_CODE]: The skill consists entirely of Markdown instructions and reference data. It does not ship with executable scripts (Python, Node.js, etc.) or binaries.
[EXTERNAL_DOWNLOADS]: No external network requests or downloads were identified. All operations are local to the skill folder.
[PROMPT_INJECTION]: No evidence of instructions attempting to bypass safety filters or override system prompts was found. The skill's purpose is actually to reinforce safety boundaries.

one-way-door