markdown-quality
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes repository Markdown files which could contain untrusted instructions.
- Ingestion points: The agent discovers and reads documentation files across the repository, including human-authored and generated content (SKILL.md).
- Boundary markers: Includes an explicit "Content boundary" directive instructing the agent to treat document text as content only and disregard embedded commands (SKILL.md).
- Capability inventory: The skill is authorized to perform file edits and execute repository-native tooling like formatters, linters, and build scripts (SKILL.md).
- Sanitization: Relies on instructional directives; no programmatic sanitization or structural validation of ingested content is specified.
- [EXTERNAL_DOWNLOADS]: The skill references well-known educational resources and official documentation for Markdown syntax.
- References the Markdown Guide for basic and extended syntax and the original Markdown design specification at Daring Fireball (references/markdown-practices.md).
- Includes a link to a community-maintained Markdown cheatsheet on GitHub (references/markdown-practices.md).
- [COMMAND_EXECUTION]: The workflow involves executing local repository tools to verify documentation quality.
- The agent is instructed to run the repository's existing formatters, linters, docs build commands, and link checkers when available (SKILL.md).
Audit Metadata