agentleFS
pobstetapobsteta/aigora/.claude/skills/research-lead/SKILL.md

research-lead

Skill0 starsChanged 46 days agoMITView on GitHub

Turn a lead (LinkedIn URL, name + company, or pasted profile) into a research package with relevant personalization and an outreach sequence, reviewed by a human before anything is sent. Run with /research-lead or ask to research a lead.

  • Reads credentials

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Test for every fact: "Does this relate to a problem they're facing that we can solve?" If no, discard it.

    line 26
  2. Very specific:

    Company research: follow assets/prompts/perplexity_research.txt with WebSearch / WebFetch (or mcp__searxng__* if configured).

    line 41
  3. Distinctive:

    Never send a DM or an email yourself.

    line 79
  4. Distinctive:

    No allowed hook: do not invent one.

    line 90

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Lead Research & Personalization (AIGORA)

What changed from the original template

The original pipeline chained five paid services (Relevance AI scraping through a workflow owned by the template author, Perplexity, OpenAI, Airtable, Google Sheets, Slack). In AIGORA the reasoning is done by you inside Claude Code, research uses WebSearch/WebFetch (or a self-hosted SearXNG MCP), storage is a local CSV, and the review card goes to kChat. Zero extra API key. The expert prompts were kept in assets/prompts/ and are your instructions for each analysis.

Objective

Produce a research package with relevant personalization only. Relevant = relates to a problem they're likely facing that we can solve. Theater = personal but irrelevant (marathons, shared schools, hobbies). Test for every fact: "Does this relate to a problem they're facing that we can solve?" If no, discard it.

Inputs

  • A LinkedIn URL, or name + company, or profile text pasted by the user.
  • context/my-business.md (replace {COMPANY_NAME} and the offer in the prompts with it).

LinkedIn pages usually block automated reading. Do not scrape LinkedIn. Use public sources (company site, press, public posts found by WebSearch) and ask the user to paste the profile text (About, Experience, recent posts) when needed.

Steps

  1. Profile: build profile_data = {full_name, first_name, company, headline, experiences: [{title, company, start, end}], recent_posts: [...], source_notes}.
  2. Company research: follow assets/prompts/perplexity_research.txt with WebSearch / WebFetch (or mcp__searxng__* if configured). Cite URLs. Output perplexity_data.
  3. Analyses (you, in order), each following its prompt file and the schemas in references/output-structures.md:
    • lead_profile (assets/prompts/lead_profile.txt)
    • pain_gain_operational (assets/prompts/pain_gain_operational.txt)
    • dm_sequence (assets/prompts/dm_sequence.txt), max 300 characters per DM
    • dm_quality_review: score 1-10 against the hard constraints of the DM prompt, approval_recommendation APPROVE / REVISE / REJECT. Revise once if below 7.
  4. Assemble .tmp/leads/<slug>/combined.json:
    {
      "linkedin_url": "...",
      "profile_data": {...},
      "perplexity_data": {...},
      "lead_profile": {"data": {...}},
      "pain_gain_operational": {"data": {...}},
      "dm_sequence": {"data": {...}},
      "dm_quality_review": {"data": {"overall_quality_score": 8, "approval_recommendation": "APPROVE"}},
      "quality_flags": []
    }
  5. Review report (HTML, for the human):
    python3 .claude/skills/research-lead/scripts/generate_review_report.py --data .tmp/leads/<slug>/combined.json
  6. Store (local CSV data/leads.csv, opens in R, LibreOffice or kSuite):
    python3 .claude/skills/research-lead/scripts/save_lead.py --data .tmp/leads/<slug>/combined.json
    Then, if Odoo is configured (ODOO_URL in .env) and the user approved the lead, offer to create the opportunity (preview first, then --confirm): python3 .claude/skills/odoo/scripts/odoo.py create-lead --from-json .tmp/leads/<slug>/combined.json [--stage ...] See the odoo skill.
  7. kChat review card (only if the user confirms): write a short Markdown summary (who, why now, chosen hook, DM1, score) and post it with python3 .claude/skills/ksuite-setup/scripts/kchat_post.py --file ....

Never send a DM or an email yourself. The human copies the approved message.

Batch

For a list of leads (CSV with linkedin_url or name,company), process them one by one, skip those already in data/leads.csv, and give a summary table at the end.

Edge cases

  • Private or unreachable profile: work from what the user pastes; note the limits.
  • Company not found: say so; use headline and posts only; lower confidence.
  • No allowed hook: do not invent one; recommend not contacting yet.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.