Docs & Merge — Car Fleet¶
Tests pass. One step left before this feature is done.
Generate the docs¶
Stay in the same chat. Run:
Read specs/brief.md, specs/roadmap.md, and the feature spec we just implemented.
Do the following in order:
1. Add docs/ to the workspace file (*.code-workspace) if not already present.
Locate the workspace file using specs/brief.md as an anchor to find the repo root.
2. Update docs/user-guide.md (create it if it doesn't exist yet).
This is a living document that grows with every feature. Structure:
If creating for the first time:
- Start with an "About this extension" section: 2-3 sentences describing
what the extension does and what problem it solves. Plain language, no AL jargon.
- Add a Table of Contents listing all feature sections (even if only one exists now).
Add a section for the feature just implemented:
- What this feature does (one paragraph)
- Setup: any one-time configuration needed before first use
- How to use: numbered steps from a user's perspective
- Related features: placeholder for now
Watch it in action¶
Review the docs¶
Before committing, ask the BC Code Intelligence MCP to check the generated docs using a subagent:
Use the Taylor Docs and Uma UX personas from BC Code Intelligence to review docs/user-guide.md in a subagent session.
Check that: the content is accurate and matches what was actually built; nothing was invented; the user steps are clear and actionable; the language is plain and avoids AL jargon.
Report any issues.
If the subagent flags anything, fix it in the same chat before committing.
Real example
Taylor found that Sonnet had written "Choose Save" in the how-to steps — but there is no Save button in Business Central. Records save automatically on navigation. One prompt fixed it across the entire section.
Commit¶
Once the agent is done, review the docs output — did it invent anything that isn't actually in the code?
Then ask the agent to commit:
Turn what you learned into a skill¶
When you close this chat and open a new one for the next feature, the agent starts fresh. Everything it learned — how you structured the docs, how to review them — is gone.
Skills fix this. Instead of re-explaining the same things in every chat, you package the learning once and invoke it by name.
Ask the agent to create a documentation skill from this session:
/create-skill
Based on what we produced in this feature — the user-guide.md structure, the review process, and the corrections made — create a skill file at .cursor/skills/generate-docs.md that automates the same documentation workflow for the next feature.
The agent will produce a skill that carries the workflow — and the lessons — into every future feature session automatically.