Update project documentation with AI
Find the instructions a release actually changes, correct their examples and leave a record of what was checked against the current behaviour.
Update project documentation with AI when a change looks smaller in the release notes than it feels to someone following the instructions. A new export format might affect a tutorial, an example command and the page that explains import limits. The docs-maintenance skill starts with the change and follows those references through the material readers actually use.
An old release note can remain an accurate account of its own version. A current tutorial promising a button that has moved needs attention. For each candidate passage, the agent checks what a person would expect to happen, whether that still matches the product and where the authoritative explanation belongs.
What it does
- Correct a getting-started example after a command, account requirement or default value changes.
- Update a guide and the links that lead to it after a documented feature moves.
- Identify a change that affects internal handover notes as well as customer-facing instructions.
- Check whether a seemingly unrelated paragraph repeats the same outdated assumption.
- Distinguish an unverified claim from an instruction tested in a suitable local environment.
Install and use
- Use the original docs-governance plugin with an authorised project. Git and Python 3 support the optional local checks. Identify the change and the version under review.
- Give the agent a known example, diff or accepted change record. Specify whether it may edit or should only report the gaps.
- Review the corrected main passage and its dependent references. Keep the observed result, test conditions and remaining questions with the handover.
FAQ
Source
Made by qshanx. Original skill: docs-maintenance, published on GitHub under the MIT licence.
View on GitHub