| name | Lead List Research |
| catalog_line | Build and refine qualified B2B lead lists: ICP patterns, fit scoring, enrichment, list cleanup. |
| description | Build and refine qualified B2B lead lists with an agent in the loop. Use when sourcing candidate accounts, turning closed-won customers into an ICP search pattern, scoring prospect fit with explained reasons, finding buyer personas, enriching contacts, or cleaning a noisy list into a repeatable outbound campaign source. Feeds the cold email outreach suite. |
| skill_type | procedure |
| altitude | domain |
| activation | progressive |
| preserve_markdown | true |
| legacy_paths | ["lead-list-research-with-claude-code","lead_list_research_with_claude_code","docs/research/youtube-library/skill-drafts/lead-list-research-with-claude-code/SKILL.md"] |
| path | apps/web/src/lib/services/agentic-chat/tools/skills/definitions/lead_list_research/SKILL.md |
Lead List Research
Identity
Use this skill to build a qualified lead list through iterative agentic research, scoring, exclusions, and contact sourcing. The goal is not to automate judgment away. The goal is to coordinate research passes, tools, and human review until the list reflects a real market opinion. The technique is tool-agnostic: it works with web search, CRM exports, enrichment tools, or spreadsheets.
This is a procedure skill at domain altitude. This skill produces the list. It does not write the outreach — downstream ownership (segment/signal grading, per-contact research) is declared in Routing.
Activation
- Build an account list from a target market or lookalike customer
- Turn closed-won customers into an ICP search pattern
- Score companies for fit and explain why they qualify
- Find buyer personas or titles for qualified accounts
- Debug noisy lists with agencies, resellers, enterprise accounts, or wrong-fit categories
- Create a repeatable outbound campaign source list
Do not use this skill when the user only needs a simple contact lookup, a generic scraping job, or a list with no qualification criteria.
Judgment
ICP Scoring Rules
- Score the account, not just the industry label.
- Require a short reason for every score.
- Treat 7+ as qualified only if the user agrees.
- Track false positives and update exclusions.
- Use homepage evidence when categories are ambiguous.
- Watch for agencies, resellers, consultants, media sites, marketplaces, wholesalers, and enterprises that exceed the intended segment.
- Do not use employee count as a universal budget proxy; some small companies are high-intent when the niche is capital-intensive.
Persona Rules
- Identify likely buyers from actual company patterns, not generic title lists.
- Split personas into tiers such as owner/founder, operator, technical lead, revenue leader, or department head.
- Account for company size: founder at small orgs, VP or director at larger orgs.
- Avoid seniority as the only selector. Titles vary by market.
- Always get at least one plausible contact per qualified account when possible.
Human In The Loop
Ask for user judgment when:
- The exclusion rule could remove valid accounts
- The market has several plausible subsegments
- The account fit depends on hidden budget or timing
- The scoring prompt returns high confidence for examples the user dislikes
- The list is being used for paid sending or high-volume outbound
