readme
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions specify the use of the
rg(ripgrep) utility to search for markdown files containing specific patterns (e.g.,rg -l '<!--' --glob '*.md' .). - [INDIRECT_PROMPT_INJECTION]: The skill is designed to parse and interpret natural language descriptions within HTML comments to determine what code constructs to sync. This allows external data within the repository to influence the agent's logic for file reading and content generation.
- Ingestion points: The skill scans all markdown files (
*.md) within the project tree, includingREADME.md,packages/*/README.md, anddocs/**/*.md. - Boundary markers: The skill uses HTML comments (e.g.,
<!-- ... -->) as markers for symbolic references. - Capability inventory: The agent performs file system searches using
rg, reads source files to extract code, and writes updated content back to markdown files. - Sanitization: There is no explicit validation or sanitization of the natural language instructions contained within the HTML comments before the agent attempts to resolve them against the codebase.
Audit Metadata