The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown instructions, stylistic guidelines, and reference examples. No malicious code or patterns were identified across any of the analyzed files.
[COMMAND_EXECUTION]: The skill requests access to standard file system tools (Read, Write, Edit, Grep, Glob) via the allowed-tools frontmatter. These tools are used for the legitimate purpose of reading, writing, and reviewing user-provided content as described in the skill's instructions.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or access to sensitive local file paths (such as .env or SSH keys) were detected.
[PROMPT_INJECTION]: The instructions are focused on establishing a specific creative voice and structure. There are no attempts to override agent safety protocols, disregard system instructions, or extract system prompts.
[REMOTE_CODE_EXECUTION]: The skill does not download external scripts, install third-party packages, or use dynamic execution functions like eval() or exec().
[SAFE]: The ingestion of user data for rewriting (Category 8) is managed through strict stylistic constraints and a quality checklist, minimizing the surface for indirect injection while maintaining the skill's primary function as a writing assistant.

reframe-voice