Systems Lab

Agent skill

bant

Audits a B2B sales call transcript against the BANT qualification framework (Budget, Authority, Need, Timeline), or sweeps a CRM export for deals missing BANT fields.

activeSelf-containedInstructions only710 words

Filed under Calls, demos and discovery.

From zime-ai/zime-gtm-skills · 41 skills · 14 · pushed 2026-08-26

What it does when it runs

Audits a B2B sales call transcript against the BANT qualification framework (Budget, Authority, Need, Timeline), or sweeps a CRM export for deals missing BANT fields. Use when qualifying an early-stage lead, deciding whether to advance a deal past a first call, or checking a pipeline export for unqualified opportunities.

Read from the skill and the 3 files bundled beside it. A skill’s own description is written to be selected by an agent, so it describes the job and not the dependencies.

Keys and connectors you must supply
None found.
Hosts it reaches
No third-party host appears in the skill or its bundled files.
Tool permissions it declares
No allowed-tools in the frontmatter. It only issues instructions, so there is nothing to bound.
Actions present in the files
None. Instructions only.

Ask about bant

Opens your assistant with this page's verified links already in the prompt.

Is this safe to install?ClaudeChatGPT
Adapt it to my stackClaudeChatGPT
What else do I need for it to workClaudeChatGPT
Rather ask a human? Talk to Cheetah
git clone --depth 1 --filter=blob:none --sparse https://github.com/zime-ai/zime-gtm-skills.git /tmp/zime-gtm-skills
git -C /tmp/zime-gtm-skills sparse-checkout set "skills/bant"
mkdir -p ~/.claude/skills/bant
cp -R "/tmp/zime-gtm-skills/skills/bant/." ~/.claude/skills/bant/

Picked up without a restart. A project skill of the same name is shadowed by your personal one. For one repository only, swap ~/.claude/skills for .claude/skills. Claude Code docs ↗

Or take the whole library

This repo ships a .claude-plugin manifest, so Claude Code can install all 41 skills at once. Plugin skills are invoked as /<plugin>:<skill>, so they never collide with your own.

/plugin marketplace add zime-ai/zime-gtm-skills
/plugin

The folder is the same in every client that implements the format — 46 of them — so if yours is not above, only the destination changes.

Reproduced in full from zime-ai/zime-gtm-skills/blob/4f134175badd08302f070c77449822c48403eeb1/skills/bant/SKILL.md, which is licensed MIT (skill frontmatter). 710 words, 11 headings.

BANT Qualification Audit

You are a sales-call qualification auditor. Your goal is to give a rep or RevOps a fast, evidence-backed advance/no-advance read on a first call.

Audits a sales conversation against the four BANT criteria. Lighter-weight than meddicc — BANT is built for a fast early-stage "should this deal advance" read, not a full late-stage qualification pass. Use meddicc instead once a deal is past first qualification and heading toward a technical or economic evaluation.

When to use this

  • A rep just finished a first call and wants a quick advance/no-advance read.
  • An SDR handed off a lead and the AE wants a structured gut-check before investing more time.
  • RevOps wants to sweep a pipeline export for deals sitting in "Qualified" that never actually had BANT covered.

Before you start

  • If .agents/gtm-context.md (or .claude/gtm-context.md) exists, read it first and don't ask for anything it already answers.
  • Run this end to end in one pass. Don't stop to ask which file, which call, or whether a borderline quote counts — decide from the transcript, note the call once in the output, and move on.
  • If the transcript is a demo, renewal, or support call rather than a first qualification call, say so in one line and still score whichever BANT criteria the conversation happens to touch.
  • If the file is truncated or a section is inaudible/unclear, score what's there and mark the affected criteria Unclear rather than guessing at what was probably said.

Modes

Transcript mode (.txt, .vtt, .json, .md)

claude "run bant on ./calls/first-call.txt"
  1. Read the whole transcript before scoring anything — a Budget number or a Timeline trigger often surfaces late, after the criterion looks Missed from the first half alone.
  2. Score each of the four criteria in references/rubric.md independently. Don't let a strong Need inflate a weak Budget, or vice versa.
  3. Run the rubric's reads-well-too check before finalizing.
  4. Write the output in the exact shape under ## Output format.

CSV mode (.csv)

claude "run bant on ./exports/pipeline.csv"

Structural hygiene sweep, not a call-quality claim — say so explicitly. For each deal row, check whether fields for Budget, Authority, Need, and Timeline are present and non-trivial (not blank, not a placeholder like "TBD"). Output a table: deal name, deal value, criteria missing, sorted by deal value descending.

Output format

For every criterion, in this order — Budget, Authority, Need, Timeline:

**Budget** — Covered
Evidence: "we have roughly $40-60k set aside for this this year"

**Authority** — Partial
Evidence: "I can move this forward but my VP signs off on anything over $20k"
Note: influencer with a real path to the decision-maker, not the final say

Then close with exactly one line:

Advance / Advance with a follow-up on [criterion] / Do not advance yet — because [the one thing that drove this call], and here's the one question that would change it: [specific follow-up].

No recap section repeating all four scores again after this — the call above is the deliverable, not a summary of it.

Do not

  • Don't average or blend the four scores into a single "qualification score" — BANT stays four separate reads, not one number.
  • Don't infer Authority from title alone ("VP of Ops" isn't automatically the economic buyer) — score it from what the transcript actually shows about their path to a decision.
  • Don't add a "recommended next email" or talk-track section unless asked — this skill's job stops at the advance/no-advance call.

Related skills

  • meddicc — once a deal is past first qualification and multiple stakeholders are in play, MEDDICC's seven letters replace BANT's four.
  • meeting-to-qualify — for turning a raw meeting transcript into a structured qualification record rather than a pass/fail read.
  • prospect-pipeline-check — for a pipeline-wide sweep earlier than BANT's CSV mode, before deals reach "Qualified."

Sample data

assets/sample-transcript.txt is a synthetic first qualification call — run the skill against it first.

What this does not do

No CRM connection, no API calls, no telemetry, no data retention beyond the current session. It reads the file you point it at and nothing else.

Files bundled with it

These load only when the skill asks for them, so they cost nothing until it runs.

Need help setting it up?

This page tells you what bant does and what it needs. Cheetah builds the agent setup it runs inside: data, CRM, sequencing and the guardrails.

Book a call →

The directory stays free. There is nothing gated behind this.