What does this skill do, and when should you use it?
Wiki Architect is a documentation-architecture skill in Microsoft’s deep-wiki plugin. It is intended for requests to create a wiki, plan documentation, map repository structure, or explain architecture at a high level. It first resolves the source-repository context, then examines the file tree, README, project technologies, and architectural layers. Its defined output is a hierarchical JSON catalogue with onboarding, getting-started, and deep-dive documentation sections.
It first runs Git commands to inspect the origin remote and current branch, and asks whether the repository is local-only or has a source URL. It then scans the repository tree and README, detects languages, frameworks, technologies, and presentation, business-logic, data-access, and infrastructure layers. It produces a JSON documentation catalogue with Contributor, Staff Engineer, Executive, and Product Manager onboarding guides, while limiting nesting to four levels, limiting sections to eight children, and requiring each prompt to reference specific files.
- A new contributor needs a progressive path from language foundations through architecture, setup, testing, navigation, and contribution practices.
- A staff or principal engineer needs an opinionated documentation plan with architecture diagrams, a domain model, tradeoffs, and a reading order.
- A team needs to map a medium or large codebase before creating a repository wiki.
- Product and executive stakeholders need audience-specific repository documentation without relying on code snippets.
How do you install this skill?
- The mandatory git commands lack defined, diagnosable handling for command failure or non-Git repositories.
- Sensitive information, private repository content, and external citation-link boundaries are unspecified.
- The output requirements are heavy and may generate more documentation than needed; no schema or minimal-output example is provided.
- The CI and test materials are repository-level evidence, not dedicated validation of this skill path.
- Shell / CLI
- Local filesystem
Git
The skill is located at .github/plugins/deep-wiki/skills/wiki-architect/SKILL.md. The README documents collection installation with npx skills add microsoft/skills, followed by selecting the needed skills in the wizard. In Copilot CLI, the documented plugin commands are /plugin marketplace add microsoft/skills and /plugin install deep-wiki@skills. The source does not document a standalone installation command for wiki-architect.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/microsoft/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/.github/plugins/deep-wiki/skills/wiki-architect" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- create a wiki for this repository,
- document this repo,
- map the codebase structure,
- create an onboarding guide.
After installing the skill or the deep-wiki plugin, use a request such as “create a wiki for this repository,” “document this repo,” “map the codebase structure,” or “create an onboarding guide.” The skill requires source-repository context to be resolved before analysis continues, including either a remote source URL or confirmation that the repository is local-only.
What are this skill's strengths and limitations?
- Covers repository scanning, architecture identification, documentation hierarchy, and onboarding design in one workflow.
- Defines four audience-specific onboarding guides for contributors, staff engineers, executives, and product managers.
- Requires file-specific citations and imposes clear catalogue-size constraints.
- Analysis cannot proceed until the source-repository context is resolved.
- The skill itself provides no test scenarios or acceptance criteria; the README’s test statistics describe the collection and do not establish coverage for this skill.
- The workflow depends on access to a repository file tree and README, while the source does not define handling for missing or unreadable files.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Wiki Architect this page ✓ Microsoft · Official | 52 · Use with care | ★ 2.7k | 3mo ago | MIT |
| LLM Documentation Index Generator ✓ Microsoft · Official | 54 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Wiki Codebase Researcher ✓ Microsoft · Official | 52 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Repository Wiki Q&A ✓ Microsoft · Official | 52 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Codex Bug Triage Assistant ✓ OpenAI · Official | 40 · Not recommended | ★ 129k | 11d ago | Apache-2.0 |
How did FollowSkills review this skill?
The skill is scoped to repository scanning and documentation generation, requires source-context resolution, and specifies local or linked citations. No malware, credential theft, or destructive default is shown, supporting a relatively strong score; however, permission boundaries, sensitive-data handling, external-link safety, user confirmation, rollback, and data-flow disclosure are incomplete, so 8 points were deducted.
Activation, sequencing, and output fields are fairly clear, and the repository contains general testing infrastructure; however, this skill has no dedicated tests or examples, failure behavior for mandatory git commands outside a Git repository is unspecified, and edge handling for scan failures, very large repositories, or missing READMEs is absent. 13 points were deducted; the static-review cap is 10.
The trigger scenarios are explicit and align with architecture understanding, wiki creation, catalogues, and onboarding documentation; however, non-fit boundaries, input limits, and semantic-conflict handling are not declared, and Chinese-language output support is not specified. The core function does not entirely depend on overseas services, so no mainland-network deduction was applied; 5 points were deducted.
The skill provides a name, description, MIT license, author, and version 1.0.0 metadata, plus activation rules, procedure, constraints, and output requirements; it lacks installation or dependency notes, parameter-stability guidance, examples, FAQs, known limitations, changelog, and explicit maintenance ownership or update path. 7 points were deducted.
The requested output includes a hierarchical catalogue, four onboarding guides, and file citations, which could directly support documentation planning; however, there is no complete schema example or representative verified output, some requirements may cause disproportionate generation, and static reading cannot establish correctness. 9 points were deducted; the static-review cap is 7.
The skill requires citations to real files, while the repository shows CI, evaluation workflows, and test code, providing limited auditability; however, the supplied material does not demonstrate skill-specific scenarios, coverage, or independently reproducible evidence for this path. 6 points were deducted; the static-review cap is 5.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →