The Agent Skills Directory

[SAFE]: The skill is a legitimate tool for sequential text editing. It includes clear instructions for version control, directory management, and task scoping, ensuring that edits are transparent and restricted to the user's provisioned workspace.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes untrusted user-supplied drafts through a multi-stage pipeline where the output of one step becomes the input for the next. This risk is inherent to text-processing tasks and is mitigated by instructions to preserve the author's voice.
Ingestion points: User-provided Draft text gathered in SKILL.md and passed to the agent network.
Boundary markers: Absent; text is passed between agents without specific delimiters or warnings to ignore embedded instructions.
Capability inventory: File system write operations (creating intermediate files and a final draft) restricted to a specific workspace directory.
Sanitization: No sanitization or validation of the input draft is specified; agents are instructed to preserve detail and voice faithfully.

editorial-pipeline