developer-docs-structure-audit skill (samber): audit how your docs are organized
Audits how a docs set is organised against the Diátaxis modes and returns a scored report with a ranked fix queue.
- Installs
- 12.7K
- on skills.sh
- Stars
- 7
- on GitHub
- Files
- 8
- besides SKILL.md
- Updated
- Nov 2026
- last repo commit
developer-docs-structure-audit makes your agent inventory every page of an existing docs set, classify each against the Diátaxis modes, find coverage gaps and navigation drift, score the set and return a queue of small moves. It does not write pages or check that documented behaviour matches the code.
It triggers on a docs structure or content audit, docs information architecture, "our documentation is a mess", or what docs are missing. It bundles one Python script, scripts/docs-inventory.py, which reads your docs folder and runs git log for each file. You need Python 3 and a docs folder or a pasted file tree.
Good for
- Docs owners whose readers cannot find what they need
- Open-source projects with no technical writer
- Teams planning a docs reorganisation or migration
Install developer-docs-structure-audit
Claude Code
- 1
Copy this prompt.
Install the developer-docs-structure-audit skill in Claude Code. Run this command: /plugin marketplace add samber/developer-relations-skills /plugin install developer-relations-skills@developer-relations-skills Then check that it loaded and tell me how to use it.
- 2
Paste it into Claude Code. It shows the change and asks before running it.
The README recommends installing the whole collection because its skills reference each other, and developer-docs-structure-audit names other skills in the collection to hand work to. The whole-collection routes add all 93 skills in the repo (counted from the git tree); the repo holds no hooks, agents, commands or MCP server files, and 14 of its skills bundle a script. This skill runs scripts/docs-inventory.py, so it needs Python 3. Pick one route per agent.
From the project README. Source.
What's in the skill
When it loads
Audits an existing developer documentation set's structure - a page-by-page inventory classified against the Diátaxis modes (tutorial, how-to, reference, explanation), mixed-mode and misplaced pages, coverage gaps per product surface, navigation drift against the file tree, a CNCF TechDocs rubric score, and a prioritized remediation queue. Use whenever the user mentions a docs structure or content audit, docs information architecture, Diátaxis, "our documentation is a mess", a docs gap analysis, what docs are missing, where a page should live, or reorganizing a documentation site - even if they only say "our docs are hard to navigate". Not for writing pages. Do NOT use for docs search ranking - use samber/developer-relations-skills@docs-seo.
Sections in SKILL.md
- Who should run this
- Interview
- Step 1 - Fix the scope and the evidence you have
- Step 2 - Build the inventory
- Step 3 - Classify every page
- Step 4 - Find what is missing
- Step 5 - Check that navigation tells the truth
- Step 6 - Score
- Step 7 - Validate before recommending a restructure
- Step 8 - Turn findings into a queue
- Step 9 - Deliver, then close the drift loop
- Failure modes
- References
Key facts
- Author
- Samuel Berthe (samber)
- Installs on skills.sh
- 12,679 (as of Nov 2, 2026)
- GitHub stars
- 8 for the whole collection repo (Nov 2, 2026)
- Last commit
- Nov 1, 2026 (whole repo)
- skills.sh audits
- Gen Agent Trust Hub pass, Socket pass, Snyk pass
- What is inside
- 9 files including SKILL.md: 6 reference files, 1 script (docs-inventory.py) and evals/evals.json
- Version in SKILL.md
- 1.0.3
- Collection size
- 93 skills in the repo (counted from the git tree)
What it includes
Asks fourteen questions one at a time; questions 1 to 4 gate the audit.
Questions 10, 12, 13 and 14 re-rank the fix queue and can delete moves and renames if you cannot ship redirects.
InterviewPrints one row per docs page and can compare the file tree with a nav config.
Columns are path, title, words, headings, max_heading_depth, code_blocks, numbered_steps, imperatives, option_table, questions, internal_links, external_links, in_nav and last_modified. For each file it runs git log -1 to get the last-modified date. The example command redirects output to inventory.tsv.
Runs code ·
python3 scripts/docs-inventory.py ./docs --nav mkdocs.yml > inventory.tsvAssigns each page a Diátaxis mode with a confidence and flags mixed-mode pages.
Tutorial, how-to, reference or explanation. If the site already uses another taxonomy it audits against that one.
Step 3 - Classify every pageBuilds a coverage matrix and checks the navigation tells the truth.
Product surfaces by the four modes. Mechanical checks cover orphans, dead nav entries, broken links, single-page sections and depth beyond three levels.
Step 4 and 5 - Gaps and navigationRates sections on the CNCF TechDocs rubric and computes a structural scorecard.
SKILL.md says the CNCF scale is published and the scorecard pass marks are its own baseline.
Step 6 - ScoreUses a card sort and tree test when possible; otherwise marks changes "proposed".
Above roughly 80% task success and 60% direct-path rate it suggests cancelling the restructure.
Step 7 - ValidateOrders five kinds of action by reader need fixed per hour.
Split a mixed page, merge duplicates, move a page, write a missing page, rename a section. It never proposes empty scaffolding or a teardown.
Step 8 - Remediation queueDelivers the report and closes with controls that stop the structure decaying.
Controls include CI-validated content types, page templates, prose and link checks, CODEOWNERS per subtree, a redirect for every moved URL and a quarterly re-audit.
Try: “our documentation is a mess”
Step 9 - Report and drift controls
From the official docs
Permissions
Medium risk
The agent runs a bundled Python script over your docs folder and calls git log for each file. Run it on a clone you trust and read the script first.
What it can do
- Read your data
- Create and change things
- Delete things
- Read-only mode
Watch out for
It runs a bundled Python script on your docs folder
docs-inventory.py imports argparse, json, os, re, subprocess and sys, reads every page it finds, and calls git log for each one with a 10-second timeout. Our advice: read the script and run it on a clone, not on a folder with secrets. github.com
The Claude app route is left out of the install box
The skill depends on a local script and a local docs folder, so the install box has no Claude app upload tab for it. github.com
It may keep notes in devrel-context.md
SKILL.md tells the agent to store a few details about the current task or project, and to use a file called devrel-context.md when its memory lives in a file. Our advice: open that file before you commit or share the folder. github.com
The plugin route installs the whole collection
The README's plugin and extension routes add every skill in the repo; its git tree holds 93. The repo has no hooks, agents, commands or MCP server files, and 14 of its skills bundle a script. Our advice: use the skills CLI command for this skill if you want only it. github.com
Security advisories
No published security advisories as of Nov 2, 2026.
Technical details
| Turning on read-only | SKILL.md describes no read-only mode. |
|---|---|
| Scopes |
|
Questions
How do I install developer-docs-structure-audit?
The README says to install every skill in the collection, because its skills reference each other. In Claude Code that is the plugin route in the install box above, which adds all 93 skills in the repo. The skills CLI can add just this one. Our advice: pick one route per agent. The inventory script needs Python 3.
What do I get at the end?
An audit report with a verdict, CNCF section ratings, a structural scorecard, a coverage matrix, findings by severity, a remediation queue and drift controls, plus a list of what the audit did not cover.
Does it need Python?
Only for the inventory script. SKILL.md says that if you cannot execute scripts or read the tree, the agent builds the inventory by hand from the navigation and pages it can open and notes in the report that it is partial.
developer-docs-structure-audit vs docs-seo?
Use this skill for how docs are organised and docs-seo for search ranking. SKILL.md says not to use it for docs search ranking, and the README describes docs-seo as on-page and technical SEO for a documentation site.
We have no writer. Is this the right tool?
It works as an interim measure. SKILL.md says CNCF's TechDocs assistance program and Google Season of Docs both give projects an external analysis, an implementation plan and an issues backlog for free, and tells the agent to recommend applying while you run this audit.
Does developer-docs-structure-audit need an API key, and where does it work?
SKILL.md names no API key, account or environment variable. The README documents installs for Claude Code, Claude.ai (add the repo as a plugin marketplace), Codex, Cursor, Gemini CLI and any Agent Skills-compatible tool through the skills CLI. This skill reads and runs things in your repository, so the install box has no Claude app upload tab for it. Our advice: use Claude Code or Codex.
Is developer-docs-structure-audit safe to install?
The folder bundles one Python script, docs-inventory.py. It reads files under the docs folder you give it, runs git log for each file, prints rows to the terminal, and the SKILL.md example redirects that output into inventory.tsv. Our advice: point it at a docs folder in a clone you trust and read the script before you run it.
Information
updated Nov 2, 2026
- Repository
- samber/developer-relations-skills
- Folder
skills/developer-docs-structure-audit
Sources
Docs, issue trackers and discussions this page is built from.
Show all 3Hide
- developer-relations-skills README · github.com
- developer-docs-structure-audit SKILL.md · github.com
- docs-inventory.py · github.com