The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses internal priority labeling (e.g., 'Rule -1', 'Rule 0') to manage stylistic trade-offs. These instructions are directed at maintaining writing quality and stylistic fidelity (such as prioritizing grammar over style or specific paragraph structures) and do not attempt to override the AI's underlying safety filters or extract system-level prompts.
[COMMAND_EXECUTION]: No executable commands or shell scripts were found. A mention of PowerShell's Get-Content is included in the documentation purely as a suggestion for users to handle UTF-8 encoding when viewing the skill files locally.
[EXTERNAL_DOWNLOADS]: The skill is self-contained within the provided markdown files. There are no external URLs, remote script fetches, or dependencies on untrusted third-party servers.
[DATA_EXFILTRATION]: There are no network-capable commands (like curl or fetch) or patterns indicating the collection of sensitive environment variables, credentials, or local files.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and transform user-provided text. While this creates a standard surface for indirect injection, the skill lacks the necessary capabilities (such as tool execution or network access) to turn such an injection into a high-severity threat. It functions as a closed-loop text transformer.

good-writing