The Agent Skills Directory

[SAFE]: The skill operates entirely within the agent's prompt context to restructure user input into professional templates. It uses standard reference files (references/examples.md, references/templates.md) and includes a Python script (scripts/validate.py) that performs local file existence and content pattern checks only. No network operations, sensitive file access, or remote dependencies were detected.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided technical notes. While this is an ingestion point for untrusted data, the instructions include a strict four-phase processing pipeline (Parse, Structure, Transform, Verify) that uses explicit boundary markers and deterministic templates. The skill's primary function is to transform 'stream-of-consciousness' or 'defensive' language into a structured business format, which acts as a form of natural language sanitization. Technical terms are preserved, but the structure is constrained by predefined templates. Evidence: Phase 1 (Extraction), Phase 2 (Categorization), and Phase 3 (Template Application) in SKILL.md.

professional-communication