An automated research harness that uses semantic search to find and rank relevant data sources for Florida county and GTM research.

Install

mkdir -p .claude/skills/exa-discovery && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/16709" && unzip -o skill.zip -d .claude/skills/exa-discovery && rm skill.zip

Installs to .claude/skills/exa-discovery

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use when running semantic web discovery for Florida county data sources, auction intel, or GTM research. Triggers on: exa search, discovery harness, find county GIS, find auction data, semantic search, county data sources, exa-discovery, exa_client, discovery pipeline, zonewise discovery, auction discovery, gtm research.
322 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)
Intermediate

Key capabilities

  • Build queries for semantic search
  • Perform Exa semantic search with livecrawl=true
  • Filter and rank search results
  • Persist findings to a Supabase discovery_results table
  • Handoff ranked URLs to Firecrawl for full content extraction

How it works

The skill builds queries, performs semantic search using the Exa API, ranks results based on relevance, and then persists them to a database, optionally extracting full content with Firecrawl.

Inputs & outputs

You give it
Mode-specific queries (zonewise, auction, gtm)
You get back
Ranked discovery results persisted to Supabase, with optional Firecrawl handoff

When to use exa-discovery

  • Finding county GIS data
  • Researching market verticals
  • Discovering auction data sources

About this skill

Exa Discovery

Role

Own semantic web discovery as evidence-grounded source ranking, not link collection theater.

Working Mode

Build queries -> Exa semantic search -> Filter and rank results -> Persist to discovery_results -> Optional Firecrawl handoff.

Focus Areas

  1. Query building -- mode-specific queries: zonewise (GIS/parcel), auction (foreclosure/clerk), gtm (market/vertical)
  2. Exa API -- semantic search with livecrawl=true, 20-25 results/query, EXA_API_KEY required
  3. Result ranking -- relevance score threshold 0.6, deduplicate by domain
  4. Supabase persistence -- discovery_results table (run migrations/20260327_discovery_results.sql first)
  5. Cost discipline -- estimate before batch runs, log token spend per query
  6. Firecrawl handoff -- pass ranked URLs to Firecrawl for full content extraction when score >= 0.8
  7. Batch mode -- all 67 FL counties at ~.38 total (within 0 cap)
  8. Dry run -- estimate cost and show query plan without executing API calls

Quality Gates

  • verify: Each result has url, relevance_score, county, mode fields
  • confirm: Cost estimate logged before any batch run (67-county ~.38)
  • check: No duplicate domains in final ranked list
  • ensure: discovery_results row persisted with source, query, score, discovered_at
  • call_out: Flag if EXA_API_KEY is missing or Exa returns 0 results for a county

Output Format

Constraints

  • NEVER run batch mode without logging estimated cost first
  • NEVER mark a source as CONFIRMED without verifying the URL returns HTTP 200
  • EXA_API_KEY must be set (GitHub secrets or env var) before any live search
  • Use node discovery/src/index.js as canonical CLI -- do not create alternatives
  • Migration dependency: migrations/20260327_discovery_results.sql must run before persistence
  • Max 25 results per query (Exa rate limits), min relevance 0.6 for inclusion

Guard Rail

Do not include discovery results with relevance_score < 0.6 in ranked output -- low-signal sources pollute downstream analysis.

When not to use it

  • When EXA_API_KEY is missing
  • When not needing semantic web discovery
  • When not searching for Florida county data sources, auction intel, or GTM research

Prerequisites

EXA_API_KEYmigrations/20260327_discovery_results.sql

Limitations

  • Max 25 results per query
  • Min relevance 0.6 for inclusion
  • NEVER run batch mode without logging estimated cost first

How it compares

This workflow provides a structured and evidence-grounded approach to semantic web discovery, including result ranking and persistence, which goes beyond simple link collection.

Compared to similar skills

exa-discovery side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
exa-discovery (this skill)04moNo flagsIntermediate
ddgs04moReviewBeginner
literature-review5592moReviewAdvanced
openalex-database488moReviewIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

ddgs

itsmeadarsh2008

Use when an agent needs to search the web, find images/news/videos/books, or extract content from a URL. Covers the ddgs Python library, CLI, and MCP server integration.

00

literature-review

K-Dense-AI

Conduct comprehensive, systematic literature reviews using multiple academic databases (PubMed, arXiv, bioRxiv, Semantic Scholar, etc.). This skill should be used when conducting systematic literature reviews, meta-analyses, research synthesis, or comprehensive literature searches across biomedical, scientific, and technical domains. Creates professionally formatted markdown documents and PDFs with verified citations in multiple citation styles (APA, Nature, Vancouver, etc.).

5591,298

openalex-database

davila7

Query and analyze scholarly literature using the OpenAlex database. This skill should be used when searching for academic papers, analyzing research trends, finding works by authors or institutions, tracking citations, discovering open access publications, or conducting bibliometric analysis across 240M+ scholarly works. Use for literature searches, research output analysis, citation analysis, and academic database queries.

48202

scientific-critical-thinking

davila7

Evaluate research rigor. Assess methodology, experimental design, statistical validity, biases, confounding, evidence quality (GRADE, Cochrane ROB), for critical analysis of scientific claims.

1888

biorxiv-database

lifangda

Efficient database search tool for bioRxiv preprint server. Use this skill when searching for life sciences preprints by keywords, authors, date ranges, or categories, retrieving paper metadata, downloading PDFs, or conducting literature reviews.

780

gpt-researcher

assafelovic

GPT Researcher is an autonomous deep research agent that conducts web and local research, producing detailed reports with citations. Use this skill when helping developers understand, extend, debug, or integrate with GPT Researcher - including adding features, understanding the architecture, working with the API, customizing research workflows, adding new retrievers, integrating MCP data sources, or troubleshooting research pipelines.

1464

Search skills

Search the agent skills registry