gong-score

Scores sales call transcripts against a rubric and returns three cited coaching fixes.

Updated Sep 11, 2026
One-click install
npx skills add https://github.com/JYeswak/grok_bot_playground --skill gong-score-jyeswak
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gong-score
Source: https://github.com/JYeswak/grok_bot_playground/tree/main/plugin/skills/gong-score
Command: npx skills add https://github.com/JYeswak/grok_bot_playground --skill gong-score-jyeswak

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Call coaching feedback tends to inflate into long lists of vague impressions that reps never act on, and scores are often given without quoting the actual transcript. This Skill scores one call transcript against a stated rubric and returns exactly three ranked fixes, each tied to a quoted transcript line. ## Core Features & Use Cases - Evidence-based scoring: Every rubric dimension gets a score, one sentence of reasoning, and a quoted transcript line; dimensions without evidence are marked unmeasured rather than scored low. - Mechanical measures: Counts talk ratio, longest monologue, questions asked, and whether a dated next step was stated, directly from the transcript. - Three-fix cap with diffing: Ships exactly three ranked fixes with replacement sentences, and diffs against the rep's prior scored call to check whether earlier fixes landed. - Use Case: A sales manager exports a call transcript each week, asks for a score against the team rubric, and receives a dated coaching note with quoted evidence and three concrete fixes to review with the rep. ## Quick Start Score this call transcript against our discovery-depth, objection-handling, and next-step rubric and give me the three fixes with quoted lines.

Frequently Asked Questions about gong-score

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I score a sales call transcript against a rubric?

Provide a speaker-labeled transcript and a rubric naming dimensions and a scale, such as discovery depth or objection handling on a 1-5 scale. The output is a dated coaching note with per-dimension scores, quoted evidence, mechanical measures, and three ranked fixes.

What happens if I do not supply a scoring rubric?

The Skill applies a stated set of default dimensions, such as discovery depth, next-step clarity, objection handling, and talk ratio, and declares those defaults explicitly in the output. It never scores against a silent or unstated rubric.

Can this tool pull transcripts directly from Gong or another call recorder?

No. There is no call-recording vendor integration; a human must export the transcript and place it where the Skill can read it. It never claims a recording API and never fabricates transcript lines.

Does the call scoring update my CRM or message the sales rep?

No. The Skill is strictly read-only: it never writes to a CRM, updates a deal, messages the rep or customer, or shares the recording. Any message to the rep is a draft that only the owner can explicitly approve for sending.

Why does a rubric dimension show as unmeasured instead of a low score?

A dimension is marked unmeasured when no transcript line can be quoted as evidence for it. The rule is that every score must cite the transcript, so missing evidence is reported as unmeasured rather than guessed as a low score.

Can I use one scored call to evaluate a rep's overall performance?

No. The Skill explicitly scores one call and stops; it never generalizes a single conversation into a judgment about a person's overall performance. Comparing two calls is supported only to check whether a specific prior fix landed.