writing-fragments
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to assist in writing by interviewing the user and storing the resulting "fragments" in a structured Markdown format. All identified behaviors align with this stated purpose.
- [COMMAND_EXECUTION]: The skill manages a local document by reading from and writing to a file path provided by the user. These file system operations are localized and standard for a productivity tool.
- [INDIRECT_PROMPT_INJECTION]: The skill re-reads the markdown file from disk before writing to preserve manual user edits. While reading external data can technically introduce a surface for indirect instructions, the skill is constrained by its design to only append or edit specific writing fragments as requested by the user, and the
disable-model-invocation: trueflag in the frontmatter further limits its ability to execute potentially harmful tool calls.
Audit Metadata