ADO Wiki Converter
Converts VitePress and GFM documentation into Azure DevOps Wiki-compatible files.
The skill requires repository-source resolution and user interaction, states that source files remain untouched, writes copies under dist/ado-wiki/, and uses only Node.js built-ins, reducing dependency and exfiltration risk. Points are deducted because it lacks explicit pre-write confirmation, output-collision handling, rollback or cleanup procedures, sensitive-data guidance, and publishing permission boundaries for Azure DevOps.
The transformation rules, sample functions, and verification checklist are fairly concrete, and the skill requires stopping when repository context is unresolved. Static evidence provides no skill-specific tests; collectMarkdownFiles, container conversion, index generation, and .order generation are not fully implemented, while link, front-matter, and Mermaid edge cases are limited. The static ceiling therefore applies.
The audience and scenario are clear: converting VitePress/GFM documentation into Azure DevOps Wiki-compatible format, with stated boundaries for same-directory links, external links, and skipped directories. Points are deducted because trigger conditions, input/output contracts, and non-fit cases are not formalized enough; Chinese-language validation and mainland-China network fit are not discussed, although the core conversion is local.
The front matter supplies a name, description, MIT license, author, and version, and the document is layered around incompatibilities, script structure, verification, and limitations. Points are deducted for missing installation and Node-version requirements, complete runnable examples, FAQ guidance, changelog, and an explicit maintenance/update path; some stated requirements also diverge from the shown main-flow pseudocode.
The core goals for Mermaid, front matter, links, and container conversion are explicit, and transformed copies are intended to be directly usable for later Wiki publishing. Points are deducted because key helper logic is omitted, while index-page generation, .order files, non-Markdown copying, and npm integration are requirements or fragments rather than a demonstrated complete implementation. Static evidence cannot establish complete usability.
The skill includes concrete transformation code, an acceptance checklist, and auditable rule descriptions, while the repository shows general CI and test infrastructure. Points are deducted because there are no skill-specific tests, real output samples, or third-party execution records, and the supplied CI/test material does not cover this skill’s key paths.
- Do not treat this SKILL.md as a verified complete build script; implement and test directory collection, index generation, .order files, non-Markdown copying, and container conversion before production use.
- Before execution, confirm the git remote, source root, and the write scope of dist/ado-wiki/ to avoid overwriting existing output; separately confirm credentials and permissions before publishing to Azure DevOps.
- Sample-check Mermaid syntax, ADO version compatibility, relative links, and Chinese-content rendering in the target Wiki.
What it does & when to use it
ADO Wiki Converter is a documentation-engineering skill in Microsoft’s Agent Skills repository. It instructs an agent to generate a Node.js ESM build script that transforms VitePress or GFM Markdown into Azure DevOps Wiki-compatible copies under dist/ado-wiki/. The workflow addresses Mermaid syntax, YAML front matter, parent-relative source links, and VitePress containers while preserving citations, diagrams, structure, and images. It is a good fit for teams migrating documentation to ADO Wiki, but it does not document publishing or authentication.
It instructs the agent to resolve the source repository context and generate scripts/build-ado-wiki.js. The script reads documentation files, strips YAML front matter, converts Mermaid fences to ADO containers, changes flowchart to graph, removes Mermaid <br> variants and long arrows, and turns parent-relative source links into plain text. It can also convert VitePress tip, warning, danger, and info containers into ADO alert blockquotes. The output goes to dist/ado-wiki/, with transformation statistics, index.md handling, .order files, and verification checks.
- A development team maintaining VitePress documentation that must also render correctly in Azure DevOps Wiki.
- A documentation owner migrating pages containing Mermaid diagrams, YAML front matter, and source-code links.
- An engineer who needs transformed copies while keeping the original Markdown untouched.
- A documentation maintainer seeking a dependency-free Node.js conversion build for CI or local use.
Pros & cons
- Provides concrete Azure DevOps Wiki fixes for Mermaid, front matter, links, and VitePress containers.
- Requires only Node.js built-in modules and no external dependencies.
- Keeps source files untouched and writes transformed copies to a separate directory.
- Includes build statistics, landing-page rules, sidebar ordering, and a verification checklist.
- The source provides instructions rather than an observed implementation or test results.
- Publishing to Azure DevOps Wiki, authentication, and deployment are outside the documented scope.
- Some Mermaid diagram types are unsupported by ADO, and inline styling may be ignored.
- The workflow requires an agent environment with shell and filesystem access.
How to install
The skill is included in Microsoft’s deep-wiki plugin. In Copilot CLI, run /plugin marketplace add microsoft/skills and then /plugin install deep-wiki@skills. The repository collection also supports npx skills add microsoft/skills, followed by selecting skills in the wizard; the source does not document a more specific single-skill selection command.
How to use
After installing the deep-wiki plugin, ask the agent: “Generate an ADO Wiki build script for this VitePress documentation, preserve citations and images, and write converted copies to dist/ado-wiki/.” The skill requires a scripts/build-ado-wiki.js script and verification of Mermaid fences, flowchart keywords, front matter, links, directory structure, copied assets, and index.md.