web-search-rag-official-quick
Perform fast, lean web research prioritising official primary sources.
Skill by derricka-lau
Chalk Labs
Need help deploying web-search-rag-official-quick?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/derricka-lau-web-search-rag-official-quick
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: ml, research
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/web-search-rag-official-quick
What's inside
- Skill name
web-search-rag-official-quick- Description
- Perform fast, lean web research prioritising official primary sources. Use only when the user explicitly invokes web-search-rag-official-quick or explicitly requests a quick, primary-source-led lookup.
- License
- MIT
- Instructions
- 54 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Web Search RAG — Official Sources Quick Mode ## Why this skill exists The user wants every response grounded in fresh web information, with a strong preference for **official and primary sources** — but fast. Don't spend 10 searches hunting down the perfect source. Instead, craft 1–2 smart queries designed to surface official sources directly, cite what you find, and move on. ## Core rules 1. **Always search first.** No message gets a response without at least one `web_search`. No exceptions. 2. **Target official sources in your queries.** Craft searches to surface primary sources directly rather than secondary coverage. 3. **Cite every factual claim.** Use system citation tags (`antml:cite`) when available, markdown links `([Source](URL))` for `web_fetch` sources. 4. **Keep it tight.** 1–2 searches for simple queries. 2–3 for moderate. Only go beyond 3 if a primary source is critically needed but elusive. Use `web_fetch` only when a search snippet is too vague to cite confidently from an official source. ## What counts as "official" (any domain, not just code) The question to ask is: "Who is the original authority on this?" Then target that source. **Tech/software:** Official docs, GitHub repos, release notes, specs. **Business:** Company IR pages, press releases, SEC filings. **Government/law:** .gov sites, legislative text, regulatory filings. **Health:** WHO, CDC, NHS, FDA, peer-reviewed journals. **Science:** Peer-reviewed papers, official datasets, patent filings. **Standards:** ISO, IEEE, IETF RFCs, W3C, NIST. **Economics:** Central banks, World Bank, IMF, OECD data portals. **Sports:** Official league sites, governing body stats. **Products:** Manufacturer spec sheets, official product pages. ## Search tactics for surfacing official sources quickly These small adjustments to your queries dramatically increase the chance of getting a primary source on the first try: - **Name the authority:** `"NHS guidelines type 2 diabetes"` not `"diabetes treatment guidelines"` - **Use official terminology:** `"ECB monetary policy decision March 2026"` not `"Europe interest rates"` - **Add document-type terms:** append `"documentation"`, `"official"`, `"release notes"`, `"specification"`, `"press release"` as appropriate - **Follow citation chains mentally:** If a snippet says "according to a WHO report," your next search should target that WHO report directly ## Recency enforcement Official sources get updated — outdated official docs are worse than current ones: - **Append the current year** to queries about evolving topics. `"CakePHP 5 migration guide 2026"` not `"CakePHP migration guide"`. - **Use recency terms** for fast-moving topics: `"latest"`, `"updated"`, `"current"`, `"March 2026"`. - **Check publication dates** on results. If the official source you found is 6+ months old on an evolving topic, spend one more search looking for a newer version — this is worth the extra query. - **Prefer the latest version** of official docs. `"Python 3.13 docs"` not `"Python docs"`. `"NHS guidelines 2026"` not `"NHS guidelines"`. Example: `"FDA guidance AI medical devices 2026"` not `"FDA AI medical devices"`. `"React 19 breaking changes latest"` not `"React breaking changes"`. ## Citation rules Place citations inline, right after the claim. When you have an official source, make its authority clear in your phrasing: "According to the official Python 3.12 docs..." or "NHS guidance states..." rather than just attaching a link. If you can only find secondary sources, say so: "According to reporting by [outlet]..." If you can't source a claim, mark it: *(unsourced — general knowledge)*. ## What not to do Don't run 5+ searches for a simple factual question. Don't `web_fetch` every result. But also don't settle for a blog post paraphrasing an official source when one more targeted search would get you the real thing. Find the balance — official source preference with minimal overhead. ## Every message, no exceptions Greetings, follow-ups, code, creative writing, thanks — all get at least one search. For trivial messages, find a quick interesting fact from an official source (a new release, a government announcement, a published dataset) and weave it in briefly.
From derricka-lau/ai-agent-contract, MIT.
More from derricka-lau/ai-agent-contract
- 0.497score
web-search-rag-quick
derricka-lau
Perform fast, lean web research with concise citations.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
python
derricka-lau
Python coding conventions — type hints, testing with pytest, code style.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
typescript-react
derricka-lau
TypeScript and React coding conventions — strict typing, functional components, testing patterns.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
web-search-rag-deep
derricka-lau
Perform exhaustive, multi-pass web research with citations.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
web-search-rag-official-deep
derricka-lau
Perform exhaustive web research prioritising official primary sources.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
php-cakephp
derricka-lau
PHP and CakePHP coding conventions — testing patterns, code style, and framework-native approaches.
SkillCFreshness: C★ 1Claude Code +9 - 0.492score
implement-change
derricka-lau
Apply a minimal, correct implementation following the approved plan.
SkillCFreshness: C★ 1Claude Code +9 - 0.462score
test-and-validate
derricka-lau
Run the strongest relevant validation sequence.
SkillCFreshness: C★ 1Claude Code +9 - 0.462score
test-strategy
derricka-lau
Plan and execute the strongest relevant test strategy for a change.
SkillCFreshness: C★ 1Claude Code +9 - 0.427score
repo-mapper
derricka-lau
Map the relevant parts of the repository before implementation.
SkillDFreshness: D★ 1Claude Code +9 - 0.427score
review-risk
derricka-lau
Evaluate a change for high-value risks before completion.
SkillDFreshness: D★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jul 21, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from derricka-lau
- 0.497score
web-search-rag-quick
derricka-lau
Perform fast, lean web research with concise citations.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
python
derricka-lau
Python coding conventions — type hints, testing with pytest, code style.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
typescript-react
derricka-lau
TypeScript and React coding conventions — strict typing, functional components, testing patterns.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
web-search-rag-deep
derricka-lau
Perform exhaustive, multi-pass web research with citations.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
web-search-rag-official-deep
derricka-lau
Perform exhaustive web research prioritising official primary sources.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
php-cakephp
derricka-lau
PHP and CakePHP coding conventions — testing patterns, code style, and framework-native approaches.
SkillCFreshness: C★ 1Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get web-search-rag-official-quick running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
web-search-rag-official-quick skill. Agent Mart. https://agentmart.dev/skills/derricka-lau-web-search-rag-official-quick (last updated 2026-10-24)