claude-md-writer
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists exclusively of markdown documentation and instructional content. There are no scripts, binaries, or automated commands included in the package.
- [SAFE]: References to system file paths, such as
~/.claude/CLAUDE.mdand/Library/Application Support/ClaudeCode/CLAUDE.md, are used correctly to explain the platform's standard memory hierarchy. - [SAFE]: The localized Russian documentation (
README.ru.md) is legitimate and contains no hidden instructions or obfuscation. - [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing project-level documentation files. Ingestion points: Reads existing
CLAUDE.mdand.claude/rules/files for refactoring. Boundary markers: None provided in the instructions. Capability inventory: The skill has no code and relies on the agent's default file system tools. Sanitization: No sanitization logic is present in the markdown instructions.
Audit Metadata