enrich-and-score
Skill0 starsChanged 14 days agoMITView on GitHub
Add Indeed employer data, LinkedIn company data and Crunchbase funding data to sourced accounts and re-score them 0-100 against the ICP, using Alpha OSINT Prospector. Use when the user says enrich these companies, which of these are funded, score my leads, rank the list, or why did this company score low. Explains every score from stored reasons and never from general knowledge.
In this repository5 other files it gives its agents.
- define-icpSkill
- export-leadsSkill
- find-decision-makersSkill
- prospectSkill
- source-accountsSkill
The file
Files
Enrich and score
Shared rules: ../prospect/references/conventions.md.
Steps
list_leads(limit=25)to see what is unenriched (enriched: false).enrich_companies(top=10)for the best unenriched accounts, or passcompany_idsfor specific ones. Worst case is about 1.5 cents per company for all three sources.- Report
matchedper source andmismatches_dropped. list_leadsagain. Scores are recomputed automatically after each stage;score_leadsis only needed after the ICP changes.
How matching works, so you can explain it
- The company website domain is the join key. The Indeed employer page runs first because it is the cheapest place to learn the website; it is accepted only on an exact normalized name match. LinkedIn lookups go by name and are accepted only when the returned website matches the account's domain, or when no domain is known on one side and the names match exactly.
- Crunchbase lookups go by domain when one is known, which is far more reliable than by name.
- A dropped match stays dropped. Tell the user they can confirm the right LinkedIn URL and you will re-run with it. Do not pick one yourself.
How the score works
fit (industry, headcount, country) + hiring (count and recency of matching roles) + funding (recency of last round) + momentum (traffic growth, urgent hiring) + reachability (named contact, email, verified email), weighted by the ICP. Each part lists its reasons in get_lead.
Known limits to state plainly when they matter:
- Round dates come from the LinkedIn funding summary. Crunchbase publicly shows round structure and investors but not dates or amounts, so "2 round(s) on record, date not public" is the honest output.
- "headcount unknown" means no source returned a size band. It is not a small company.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

