effective-web-research

Make every external web lookup effective and credible, and turn "deep research" requests into auditable, source-evaluated reports. Grounded in two established frameworks — CRAAP (decades of information-literacy science) and Google E-E-A-T (the web's dominant content-quality rubric) — so the rules are defensible, not opinionated.

This skill is about the external web. It does not compete with internal code-search tools.

When to use / When NOT to use

Use when the agent is about to (or is asked to) gather information from the external web — official docs sites, third-party library docs, RFCs, reputable blogs, papers, real-world data.

Do NOT use for:

• Reading files inside the workspace/repo (read, grep, ast-grep)
• Navigating the local codebase (LSP, goto_definition, explore agent)
• Reading installed dependencies in node_modules / venv / site-packages
• Creating or editing documents

Those are internal search tasks with different tools and different quality bars.