The Agent Skills Directory

[SAFE]: The skill operates entirely within a localized environment. It manages configuration files in ./writing-workspace/styles/ and material data in ./writing-workspace/materials/, which are standard functional behaviors for a writing tool.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a data consumer for untrusted inputs such as user-provided reference articles and the contents of a local material library.
Ingestion points: User-provided themes and reference articles, as well as the ./writing-workspace/materials/index.jsonl file.
Boundary markers: The skill does not implement specific delimiters or safety instructions to separate ingested data from the core prompt logic.
Capability inventory: The skill's capabilities are restricted to reading local files and writing generated text to the ./writing-workspace/drafts/ directory. It does not have access to network tools, shell command execution, or dynamic code evaluation.
Sanitization: No explicit content filtering or validation for the reference articles was identified.

compose