The Agent Skills Directory

[SAFE]: The skill is composed entirely of markdown instructions and does not include any executable scripts, binary files, or configuration templates.
[COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or instances of dynamic context injection (e.g., the !command syntax) within the skill file.
[EXTERNAL_DOWNLOADS]: The skill instructions explicitly state it retrieves nothing and cites nothing. No URLs for remote code or external package registries were detected.
[PROMPT_INJECTION]: The instructions do not contain patterns aimed at overriding safety filters or extracting system prompts. It establishes strict internal rules to ensure the agent only maps hypotheses rather than asserting facts.

research-scoping