skills.rest
dbt-labs
dbt-labsOFFICIAL
Explore all skills by this author →
WebsiteGitHub

maintaining-dbt-documentation

Audits dbt documentation coverage and drafts missing model and column descriptions in the project's existing style.

699|61|Updated Jan 8, 2026
Search Tags:#audit#YAML#coverage#analytics engineering#Documentation#DBT#Manifest
One-click install
npx skills add https://github.com/dbt-labs/dbt-agent-skills --skill maintaining-dbt-documentation
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: maintaining-dbt-documentation
Source: https://github.com/dbt-labs/dbt-agent-skills/tree/main/skills/dbt/skills/maintaining-dbt-documentation
Command: npx skills add https://github.com/dbt-labs/dbt-agent-skills --skill maintaining-dbt-documentation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

As dbt projects grow with many contributors, new models routinely land without YAML documentation, and coverage drifts below 100%. Manually finding every undocumented model and writing descriptions that match the project's conventions is slow and inconsistent.

Core Features & Use Cases

  • Coverage Auditing: Runs dbt parse and a Python script against target/manifest.json to report per-folder model and column documentation coverage, regardless of YAML layout or {% docs %} blocks.
  • Style-Matched Drafting: Reads already-documented models to mirror the project's YAML layout, description mechanism, and shape before writing new entries.
  • Folder-Scoped Backfill: Documents one folder at a time, validates with dbt parse, and hands back a reviewable git diff without committing.
  • Use Case: After a sprint where several contributors added staging and mart models, run the audit to find the gap, draft descriptions traced through ref()/source() to their origins, and review the diff before merging.

Quick Start

Audit my dbt project for undocumented models and draft missing descriptions for the marts folder in our existing documentation style.

Frequently Asked Questions about maintaining-dbt-documentation

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

FAQPage Schema
How do I find undocumented models in my dbt project?▼

Run dbt parse to regenerate target/manifest.json, then run the audit_coverage.py script from the project root. It reports per-folder model and column documentation coverage and lists undocumented models for any folder you pass as an argument.

How to backfill missing dbt model descriptions at scale?▼

Audit coverage first, then document one folder at a time by reading each model's SQL, tracing columns through ref() and source() to their origins, and mirroring the YAML style of already-documented models. Validate with dbt parse and review the git diff before committing.

Does the dbt documentation audit require a warehouse connection?▼

No. The audit reads target/manifest.json produced by dbt parse, which needs no warehouse connection. It counts only columns declared in YAML; surfacing undeclared warehouse columns requires dbt docs generate and inspecting target/catalog.json separately.

Why does the dbt doc coverage audit show stale results?▼

The audit is only as fresh as target/manifest.json. If you edited YAML or SQL since the last parse, rerun dbt parse before running audit_coverage.py so the manifest reflects the current project state.

Will this skill commit documentation changes automatically?▼

No. It writes YAML entries, validates them with dbt parse, and shows the diff via git diff for human review. It never commits or pushes unless the user explicitly asks.

Related Skills

Explore More Data & Analytics Skills→
mporentamporenta

lineage-analyzer

Map dbt model dependencies and generate lineage visualizations.

Community
Advanced
AltimateAIAltimateAI
803

dbt-pr-review

Reviews dbt pull requests with deterministic lineage, equivalence, and PII checks.

Official
Intermediate
davekim917davekim917
1

analytics-engineering

Standardize dbt project structure, naming, tests, and documentation.

Community
Advanced
Avila-TekAvila-Tek

domain-model-generator

Generate and update docs/domain_model.md from project context and logs.

Official
Advanced