The Agent Skills Directory

[SAFE]: The skill is comprised of Markdown documentation, templates, and reference materials. There are no executable scripts (Python, Node.js, Shell, etc.) provided within the skill package.
[SAFE]: The skill uses clear, instructional language to guide the AI in documentation tasks such as creating, updating, and reviewing README files. No prompt injection or bypass attempts were detected.
[SAFE]: While the templates/internal.md file contains placeholders for sensitive information like DATABASE_URL and API_KEY, these are clearly intended as templates for the user's own documentation. The instructions explicitly advise users to store actual secrets in secure locations like 1Password or Vault, promoting security best practices.
[SAFE]: The skill includes several reference documents (e.g., 'Art of README', 'Standard README Specification') that link to well-known and trusted open-source resources, package registries (NPM), and documentation tools (Docusaurus, Read the Docs). All external links are legitimate and relevant to the skill's primary purpose.
[SAFE]: The skill outlines a workflow where the agent reads project metadata (e.g., package.json) and source files to ensure documentation accuracy. Since the skill does not request any tool permissions (e.g., network access or file writing) and only generates text responses for user review, there is no exploitable attack surface for indirect prompt injection or data exfiltration.

crafting-effective-readmes