agentleFS
mardab96mardab96/b2b-lead-generation-claude-skills/lead-scoring-sanity-check/SKILL.md

lead-scoring-sanity-check

Skill5 starsChanged 4 months agoMITView on GitHub

Tests B2B lead scoring rules against real CRM outcomes, sales feedback, qualification stages and revenue signals. Use when high scores stop predicting closed deals, or before letting scoring decide routing.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 11 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    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. Very specific:

    Use the shared quality bar in ../references/output-standard.md and ../references/skill-design-principles.md when those files are available.

    line 8
  2. Distinctive:

    Do not use this skill for broad lead-generation advice without source, CRM, sales or qualification evidence.

    line 16
  3. Distinctive:

    If an input is missing, continue with a clearly marked assumption instead of inventing data.

    line 24
  4. Distinctive:

    Use sales rejection reasons to identify misleading scoring signals.

    line 31
  5. Distinctive:

    Use this table when the user provides scoring and CRM outcome data:

    line 36
  6. Distinctive:

    If the data does not connect to revenue, pipeline, qualified leads or conversion quality, label the recommendation as a hypothesis.

    line 50
  7. Distinctive:

    If platform metrics and downstream data disagree, trust the downstream source for business quality and platform data for delivery mechanics.

    line 51
  8. Distinctive:

    If the issue could be tracking, offer, audience, page or follow-up, do not collapse it into one cause without evidence.

    line 52
  9. Distinctive:

    Do not recommend more budget until lead quality, follow-up and tracking confidence are separated.

    line 53
  10. Distinctive:

    Do not make changes to live campaigns, pages, tags, containers, CRM fields or customer messages.

    line 75
  11. Distinctive:

    Do not claim performance impact without evidence.

    line 76
  12. Distinctive:

    Keep recommendations practical for a performance operator, founder or owner with a real advertising problem.

    line 78

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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 Scoring Sanity Check

Use the shared quality bar in ../references/output-standard.md and ../references/skill-design-principles.md when those files are available.

Use this skill when

  • the user shares lead source, CRM stage, sales note, form, landing page or campaign data tied to lead scoring sanity check.
  • the next decision could change targeting, qualification, scoring, follow-up, sales handoff or budget.
  • lead volume looks acceptable but SQL, opportunity, closed-won, rejection or response-speed data raises doubt.

Do not use this skill for broad lead-generation advice without source, CRM, sales or qualification evidence. Use it when a real B2B lead quality decision is on the table.

Required input

  • business model, ICP, offer, ACV or deal value range, sales cycle and main conversion goal.
  • ad, landing page, lead form, CRM, call note, email or campaign data relevant to this diagnostic.
  • time window, traffic source, lead volume and downstream outcomes where available.
  • what decision the user is trying to make next: create, fix, scale, pause, brief sales or investigate.
  • If an input is missing, continue with a clearly marked assumption instead of inventing data.

Analysis workflow

  1. Inventory current scoring factors, point values, thresholds and MQL/SQL handoff rules.
  2. Compare score buckets with MQL, SQL, opportunity, closed-won, closed-lost and disqualified outcomes.
  3. Calculate or estimate false positives, false negatives, source bias, firmographic bias and activity bias.
  4. Use sales rejection reasons to identify misleading scoring signals.
  5. Recommend keep, recalibrate, split scoring by segment, remove weak factors or add missing revenue-linked signals.

Diagnostic rubric

Use this table when the user provides scoring and CRM outcome data:

Score bucketLead volumeMQL rateSQL rateOpportunity rateClosed-won rateRejection patternScoring verdict
Score range or lifecycle tierCount and source mixPercent or missing-data markerPercent or missing-data markerPercent or missing-data markerPercent or missing-data markerCommon sales reason or disqualification reasonKeep / lower / raise / split / remove factor

Then call out:

  • False positives: high-score leads rejected by sales or stuck before opportunity.
  • False negatives: low-score leads that become SQLs, opportunities or closed-won.
  • Bias source: activity, source, firmographic, title, company size or form type.

Decision rules

  • If the data does not connect to revenue, pipeline, qualified leads or conversion quality, label the recommendation as a hypothesis.
  • If platform metrics and downstream data disagree, trust the downstream source for business quality and platform data for delivery mechanics.
  • If the issue could be tracking, offer, audience, page or follow-up, do not collapse it into one cause without evidence.
  • Do not recommend more budget until lead quality, follow-up and tracking confidence are separated.

Output format

FindingEvidenceLead quality impactRecommended actionConfidence
Specific diagnostic claimData, screenshot, report, note or missing-data markerBusiness or signal consequenceSmallest useful next step and ownerHigh / Medium / Low

End with:

  • Decision: fix / test / monitor / ask for data / do not act yet
  • Approval needed: yes/no and what would change if approved
  • Missing data: only the inputs that would materially change the recommendation

Practical example

User: "Here are CRM stages, source data and sales notes for lead scoring sanity check. What should we change before the next campaign move?"

Assistant should: test each scoring rule against real outcomes, name the rules that predict nothing, and stop at a revised scoring model for approval.

Guardrails

  • Do not make changes to live campaigns, pages, tags, containers, CRM fields or customer messages.
  • Do not claim performance impact without evidence.
  • Mark missing data clearly.
  • Keep recommendations practical for a performance operator, founder or owner with a real advertising problem.

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.