write-root-readme
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes repository source code and metadata to generate a new README. This creates a surface for indirect prompt injection where malicious instructions hidden in the repository's content (such as code comments) could influence the agent's behavior or output.
- Ingestion points: Project source files, example-readme.md, and repository structure metadata.
- Boundary markers: No specific delimiters or "ignore instructions" directives are provided to separate repository content from agent instructions.
- Capability inventory: File system write access (README.md) and shell command execution (rumdl).
- Sanitization: No explicit sanitization or filtering of repository content is mentioned before processing.
- [COMMAND_EXECUTION]: The instructions require the agent to run the
rumdl fmt .command in the shell to validate the markdown format of the repository root. This invokes a subprocess call on the host environment. - [EXTERNAL_DOWNLOADS]: The
example-readme.mdtemplate references external assets including badges and images from well-known services such as Cloudflare, GitHub, Shields.io, and Socialify. These references are standard for repository documentation and originate from established sources.
Audit Metadata