mcpField
S
search

search

github.com/htlin222Productivity7840 installsUpdated 3w ago
Repository

About search

Search is a research skill for finding and synthesizing information across deep research, fact-checking, competitive analysis, documentation lookup, and trend analysis. It combines query strategies with a workflow for scoping questions, following up on sources, and verifying claims.

Key features

  1. 01Formulate targeted search queries
  2. 02Filter searches by domain
  3. 03Define research scope and sources
  4. 04Follow up on relevant results
  5. 05Cross-check facts and source credibility

Use cases

  1. 01Researching a topic from authoritative sources
  2. 02Verifying a factual claim across sources
  3. 03Comparing competitors or tracking trends

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/htlin222/dotfiles --skill search

After installing, type/searchin Claude Code to confirm it loaded.

Details

From the project README

Search Research

Find and synthesize information effectively.

When to use

  • Deep research tasks
  • Fact verification
  • Competitive analysis
  • Documentation lookup
  • Trend analysis

Search strategies

Query formulation
# Exact phrase
"error handling best practices"

# Exclude terms
python tutorial -beginner

# Site-specific
site:github.com react hooks

# File type
filetype:pdf security audit

# Date range (use WebSearch date filtering)
react 18 features 2024
Domain filtering
# WebSearch with domain filtering
WebSearch(
    query="kubernetes best practices",
    allowed_domains=["kubernetes.io", "cloud.google.com", "docs.aws.amazon.com"],
)

# Exclude unreliable sources
WebSearch(
    query="health benefits of X",
    blocked_domains=["pinterest.com", "quora.com"],
)

Research workflow

1. Scope definition
  • What specific question needs answering?
  • What type of sources are authoritative?
  • What time period is relevant?
2. Initial search
# Broad search first
queries = [
    "main topic overview",
    "main topic best practices",
    "main topic common problems",
]
3. Deep dive
# Follow up on promising results
for result in initial_results:
    if is_authoritative(result):
        content = WebFetch(url=result.url, prompt="Extract key findings")
        facts.append(content)
4. Verification
  • Cross-reference claims across sources
  • Check publication dates
  • Verify author credentials
  • Look for primary sources

Output format

## Research Summary: [Topic]

### Key Findings

1. **Finding 1** - [Source](url)
   - Supporting detail
   - Supporting detail

2. **Finding 2** - [Source](url)
   - Supporting detail

### Consensus

- Points that multiple sources agree on

### Contradictions

- Areas where sources disagree

### Gaps

- Questions that couldn't be answered

### Sources

- [Title](url) - Credibility: High/Medium/Low

Source evaluation

IndicatorHigh CredibilityLow Credibility
Domain.gov, .edu, major publicationsUnknown, user-generated
AuthorNamed expert, organizationAnonymous, unclear
DateRecent, regularly updatedOutdated, no date
CitationsLinks to sourcesNo references
BiasBalanced, factualPromotional, extreme

Examples

Input: "Research best auth solutions for SaaS" Action: Search auth providers, compare features, check reviews, summarize

Input: "Verify this claim about performance" Action: Find benchmarks, check methodology, cross-reference results

Comments

Markdown supported