crafting-readme-files

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill contains multiple markdown code blocks featuring curl | bash command patterns. These are documented as installation templates for developers to include in their own project documentation (e.g., using placeholders like https://raw.githubusercontent.com/user/repo/main/install.sh). They are purely instructional and do not represent executable code within the skill package.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides a prompt template for the agent to read and revise external README.md files, which serves as a potential ingestion point for untrusted data.
  • Ingestion points: The prompt template in SKILL.md instructs the agent to read an existing README file.
  • Boundary markers: No specific delimiters or "ignore instructions" markers are defined in the prompt template to isolate the input content.
  • Capability inventory: None. The skill is composed entirely of markdown documentation and does not define any executable scripts, tool usage, or network capabilities.
  • Sanitization: No sanitization or validation of the input README content is suggested.
  • [SAFE]: All identified patterns are contextualized as standard documentation best practices. The skill's metadata accurately reflects its functionality, and the author uses their own verified repositories as reference examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 11:30 PM