agentleFS
andrerodriguesopsandrerodriguesops/gtm-engine/.claude/skills/sales-intake/SKILL.md

sales-intake

Skill0 starsChanged 10 days agoMITView on GitHub

The single "drop it here after a call" entry point. Use when someone drops a call transcript, fills INTAKE-FORM.md, or says "process this call", "log and score this call", "I just got off a call with X", "update the deal from this transcript". Runs the post-call flow: match the deal, score via sales-call-scorer, write the CRM deal (log entry, next step, blockers, roles, the five last_call_* fields), tier the call, create tasks, draft the follow-up email. Runs interactively, or unattended from the postcall routine.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Search active deals in the pipeline named in crm-records.

    line 39
  2. Very specific:

    Dedup: If the log already has an entry dated to this call, stop: it was processed.

    line 43
  3. Very specific:

    Call sales-call-scorer with the transcript and the deal's stage mapped to a call type (crm-records stage table).

    line 49
  4. Very specific:

    If crm-records records card_route: slack-connector (a CRM plan with no workflows), post the card yourself through the Slack connector right …

    line 73

≈ “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

Sales intake — post-call

Self-contained: the CRM write lives here. The only other skill it calls is sales-call-scorer.

Scope. Starts after the deal exists. Never creates a deal (one creator per object; see crm-records). Never moves a stage. Never sends an email.

Two modes. Interactive: a person pastes a transcript and confirms each write. Unattended: agents/postcall/RUNBOOK.md runs this with no one watching. Confirmations are waived; anything that would have been a question becomes a line on the card and a CRM task for the deal owner.

Plain text only in every output. No emojis.

Where the record lives. The CRM deal's description log is the record of every call. A clients/<slug>/ folder starts at proposal, when there are documents to keep (close/CLOSE.md).

Step 0 — Is it a call to process?

At least one attendee outside your domain, and a transcript of more than a few hundred words. Internal meetings are skipped. In interactive mode a transcript a person pasted is always processed; the length floor applies only to unattended runs. In unattended mode, the runbook also checks the transcript is finished (see its stub rules).

Step 1 — Match the deal

  1. Company from the transcript and attendee email domains.
  2. Search active deals in the pipeline named in crm-records.
  3. One match → go on. Several → ask (unattended: task for a person, stop). None → stop. Do not create a deal. Say which deal is missing and who should create it.
  4. Pull the deal, its contacts with buying roles, and its full description log.
  5. Dedup. If the log already has an entry dated to this call, stop: it was processed. hs_next_step and blockers are replaced, not appended, so re-running an old call overwrites current state with stale state.

Step 2 — Score

Call sales-call-scorer with the transcript and the deal's stage mapped to a call type (crm-records stage table). Keep the gaps; they become tasks.

Step 3 — Tier

TierTriggerEffect
same_dayA buying signal from PLAYBOOK.md § Buying signalsCard names the person who acts today. Tasks due today
escalateGrade D or F, sentiment at-risk or critical, or DEAL OVER-ADVANCEDCard leads with the recommended move
standardEverything elseRecap, tasks, follow-up draft

same_day beats escalate when both fire. The buying signal is the thing to act on.

Step 4 — Write the deal

From the callFieldRule
Agreed next actionhs_next_stepReplace. Owner and date or it is not a next step
What happeneddescriptionAppend a dated entry (crm-records format) with the Call score: line
Current obstaclesblockersReplace. Lead with the scorecard's "MEDDPICC open" letters, then the obstacle in one line each (E: CFO not met)
Stakeholder rolescontact hs_buying_roleSet when named or confirmed
Stage readiness—Propose only, in the recap
The call, for the cardthe five last_call_* fieldsWrite all five in the same update. last_call_at fires the Slack card

If crm-records records card_route: slack-connector (a CRM plan with no workflows), post the card yourself through the Slack connector right after the write, with the fields the workflow would show: deal, grade, score, tier, headline, next step, blockers, owner, record link.

Interactive: show the proposed writes and confirm. Unattended: write.

Step 5 — Tasks

One CRM task per scorecard gap not already covered by an open task. Owner is the deal owner (unowned deal → default_owner in crm-records), due date set. Every task has an owner and a date or it is not a task.

Step 6 — Follow-up draft

A draft in the rep's email, replying in the existing thread when there is one: what was agreed, in the client's own words from the transcript, and the next step with its date. The rep sends it. No invented claims; capabilities only from product-context, prices only from commercial.md.

Step 7 — Recap

[company] — [call type] — [grade] [score]/100 — tier [tier]
NEXT STEP: [owner, action, date]
BLOCKERS: ...
STAGE: stays [X] | propose [Y] because ...
TASKS CREATED: ...
FOLLOW-UP: drafted in [thread]

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.