// Gather, improve, and curate repository documentation and wiki content. Use when asked to update, analyze, or modify WIKI instructions/documentation, or when asked to pull an external wiki snapshot into the repo and curate repo-relevant knowledge into a committed, searchable knowledge base under `.ai/wiki/`.
| name | repo-documentation-gatherer |
| description | Gather, improve, and curate repository documentation and wiki content. Use when asked to update, analyze, or modify WIKI instructions/documentation, or when asked to pull an external wiki snapshot into the repo and curate repo-relevant knowledge into a committed, searchable knowledge base under `.ai/wiki/`. |
Keep the external wiki as the source of truth, but make its repo-relevant knowledge usable inside this code-repo workspace by maintaining:
.ai/Temp/wiki/ (not committed).ai/wiki/ (committed, AI-managed)The repo-local KB must copy relevant content into .ai/wiki/ so it is indexable by repo tooling (RAG). Do not rely on links back to the external wiki for the actual guidance content.
Expectation: curate broadly across the snapshot.
Required: build an explicit review checklist of all snapshot markdown pages (.ai/Temp/wiki/**/*.md) and work through it; do not skim only a small subset.
When asked to update wiki documentation, do this without asking questions:
Refresh snapshot: run .ai/skills/repo-documentation-gatherer/tools/PullWikiIntoTemp.ps1 to materialize the wiki under .ai/Temp/wiki/.
Inventory the snapshot (required): generate a complete list of snapshot markdown pages and use it as the checklist.
Command (PowerShell):
Get-ChildItem -Path .ai/Temp/wiki -Recurse -Filter *.md -File | Select-Object -ExpandProperty FullName
.order files (when present) to follow intended reading order.Curate: mirror a minimal subset of the wiki structure under .ai/wiki/ and copy repo-relevant content into matching KB markdown files.
Do not ask clarifying questions about scope, target pages, or desired outcomes. Assume the task is: refresh the snapshot and curate repo-relevant content into the KB.
Only ask questions if the workflow is blocked (e.g., missing git credentials/access, missing script, or the snapshot cannot be pulled).
Treat content as repo-relevant if it enables work in this repository.
A section is repo-relevant if it answers any of these:
The parent wiki may contain non-coding content (process, product, domain). Curate it only when it affects working with, operating, or changing this repo.
Extract, in order:
Do not extract secrets, credentials, or customer-sensitive data.
The KB must contain .ai/wiki/index.md as the entry point.
.ai/wiki/ (e.g., Illuminate/, Coding-Guidelines/, Accessibility/)..ai/wiki/index.md as a short table-of-contents linking to the curated KB pages.Write short, actionable guidance (checklists and commands). Prefer distillation, but keep the guidance content itself inside .ai/wiki/ (copy from the snapshot and distill there). Do not link back to the external wiki for long details; instead, copy any repo-relevant details that must be searchable.
Keep .ai/wiki/ in sync with the snapshot: if content is removed or materially changed in the external wiki snapshot, remove/update it in .ai/wiki/ in the same change.
Do not add any repo-derived facts, assumptions, or helpful background that is not explicitly present in the external wiki snapshot; when unsure, omit rather than invent.
Update, create, improve, and synchronise this repository's AI agent instructions and related assets (including skills). Use when the user asks to create or edit a skill/SKILL.md, modify the agent's own instructions/processes, restructure instruction governance, migrate instruction content into skills, or run/adjust the sync pipeline that publishes `.ai/` sources into agent-specific folders. Load this skill before writing any SKILL.md, .instructions.md, or touching any skills/ folder (.ai/, .claude/, .roo/, .github/). It tells you the correct location (.ai/) and the sync step, so files end up in the right place.
Rasterize specific mermaid blocks inside a markdown file into PNG images. Use when mermaid diagrams don't render in the target viewer (GitHub, Notion, PDF, etc.). Triggers on: rasterize diagrams, mermaid isn't rendering, convert mermaid to images, replace mermaid blocks with images.
AI-assisted pull request review workflow for Azure DevOps Git repositories. Use when the user asks to review a pull request, perform a PR review, analyse PR changes, or run the PR review pipeline. Provides a complete scripted workflow to fetch PR metadata, prepare a local branch folder, export diffs, generate review reports and checklists, and post review comments back to Azure DevOps. Requires Python 3.9+ and Git.
Create and maintain automated tests in Microsoft-native/.NET projects with a minimal stack — MSTest runner, `System.Windows.Automation` for Windows desktop, Playwright for real browser smoke. Code-first, single-source-of-truth. Use whenever the user wants to write, review, structure, migrate, or plan tests for any .NET project (unit, API, database, WPF, WinUI, MAUI, Blazor, Razor) — even when the request says "test" without naming a framework.
Generate or refresh `.ai/repository-analysis.instructions.md` whenever the user needs a repository-wide map of architecture, projects, technology stack, developer workflows, CI/testing, documentation, or Mermaid diagrams. Use this skill even when the request only names one slice of that work—such as “summarize the repo”, “map dependencies”, “document the stack”, “refresh onboarding context”, or “update architecture notes”—because those tasks usually need the same factual analysis workflow.
Establish and enforce deterministic path patterns across Code ↔ UI route/menu ↔ Test for any project. Use this skill before creating, renaming, or relocating any UI component, page, view, view-model, route, controller, desktop panel, or test file — it tells you the exact expected path in each of the three columns before you touch disk. Also use when onboarding a repo (generates `.ai/solution-patterns.instructions.md` + `.ai/solution-patterns.csv`), when auditing a project for structural consistency, when asked "where does X live?", or whenever the user mentions folder layout, route mapping, breadcrumbs, menu structure, project structure, naming conventions, solution structure, or "which file is the UI for this route". Load this skill before modifying any coding project — if you skip it you will invent paths instead of deriving them, and you will drift from the project's declared convention.