format-github-markdown
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill configuration references a JSON schema hosted on GitHub's official content delivery network (raw.githubusercontent.com). This is a standard reference for tool configuration and does not include executable code.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes and modifies Markdown files provided by the user (ingestion points). While this creates a surface for untrusted content, the instructions contain strict boundary markers requiring the agent to only adjust formatting and syntax, explicitly prohibiting changes to facts, commands, or wording. The scripts
markdown_toc.pyandreflow_markdown.pyprovide the file-write capability, and no sanitization of the input markdown is performed.\n- [COMMAND_EXECUTION]: The skill utilizes local Python scripts for formatting tasks. These scripts are self-contained, do not perform network operations, and are restricted to processing the files provided as arguments.
Audit Metadata