kb-gap-mining

Clusters resolved Plain support tickets into topics and drafts missing knowledge-base articles as pull requests.

20.2k|3.4k|Updated Oct 5, 2024
One-click install
npx skills add https://github.com/kortix-ai/suna --skill kb-gap-mining
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kb-gap-mining
Source: https://github.com/kortix-ai/suna/tree/main/packages/starter/templates/marketplace/runtime/skills/kb-gap-mining
Command: npx skills add https://github.com/kortix-ai/suna --skill kb-gap-mining

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Support teams repeatedly answer the same questions in tickets while the knowledge base stays silent on those topics. This Skill closes that loop by mining resolved Plain threads on a weekly schedule, finding recurring topics with no existing article, and drafting new KB content from the real resolutions — without ever publishing or merging on its own.

Core Features & Use Cases

  • Ticket-to-Topic Clustering: Pulls recently resolved Plain threads and groups them by underlying intent, so differently worded tickets about the same problem count as one topic.
  • Gap Detection with Deduplication: Cross-references each topic against the KB repo's existing articles and a durable ledger of previously drafted or below-threshold topics, so nothing gets redrafted or duplicated.
  • Draft-Only PR Workflow: Writes articles from actual ticket resolutions (with customer details stripped), commits them on an isolated branch, and opens a single labeled PR per run for human review — never merging or publishing.
  • Use Case: A support team resolves dozens of password-reset tickets each week but has no KB article on it. The weekly run detects the volume, confirms no article exists, drafts one from the real resolution steps, and opens a PR titled "kb: draft 2025-W32 batch" for an editor to review.

Quick Start

Ask the agent to run a KB gap mining pass over the past week's resolved Plain tickets and open a draft PR with any missing articles.

Frequently Asked Questions about kb-gap-mining

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

FAQPage Schema
How do I turn support tickets into knowledge base articles automatically?

This Skill pulls resolved Plain threads, clusters them into recurring topics by intent, and drafts KB articles from the actual resolutions used in those tickets. It opens one draft PR per weekly run so a human reviews and publishes.

How does the skill avoid drafting duplicate KB articles?

Each topic is checked against the existing articles in the KB repo path and a durable ledger file recording covered, drafted-pending-review, and below-threshold topics. Topics already covered or sitting in an open PR are skipped.

Can the agent publish or merge KB articles itself?

No. The agent only creates an isolated branch and opens a draft pull request labeled kb-gap. Publishing and merging are strictly reserved for human reviewers, and the agent never pushes to the live docs branch.

Does ticket mining expose customer data in drafted articles?

No. All customer names, account identifiers, and ticket-specific details are stripped before anything is written to the KB path. Only the generalized question and resolution steps appear in the draft.

What happens when a topic has too few tickets to justify an article?

Low-volume topics are logged in the ledger as below-bar rather than drafted. If a topic later crosses the volume threshold in a future run, it gets promoted into that week's draft batch.