The Agent Skills Directory

[SAFE]: The skill defines a set of instructions for professional prose editing. It focuses on style improvements such as removing defensive disclaimers and stacked hedges while explicitly requiring the preservation of meaningful facts, limitations, and legal or safety context.
[PROMPT_INJECTION]: While the skill instructs the agent to reduce "self-protective narration" and "wording the assistant adds," these instructions are strictly scoped to the stylistic quality of the reader-facing prose being edited. The instructions do not attempt to bypass core safety filters or override ethical guidelines.
[INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface as it processes user-provided material for its rewrite and audit functions (SKILL.md). While it lacks explicit boundary markers or input sanitization, the skill contains no exploitable capabilities (such as network access, file writing, or command execution). It is limited to generating text output for the user, rendering the surface functionally inert.

stss