- Home
- Skills
- Technical Documentation
- Repo Atlas: Auto Wiki Blueprint
Works with the AI Agents you already use
Repo Atlas: Auto Wiki Blueprint
Turn any repository into an evidence-linked wiki blueprint: layered docs map, senior and beginner onboarding paths, and page-generation prompts.
Free
Repo Atlas: Auto Wiki Blueprint
Example session with this skill installed
I just inherited a Go monorepo at ~/work/billing-platform: three services under cmd/, an internal/ tree nobody fully explains, a stale README, and CI in .github/workflows/. Build me a full documentation atlas before the next onboarding round. I need an orientation section, separate onboarding tracks for senior engineers and juniors, a subsystem breakdown grounded in the real directories, a getting-started path, and a decision-records branch for obvious tradeoffs. Output the catalogue as nested JSON with cited source paths per topic and a concrete generation prompt for every page, so our internal wiki can be built directly from it next sprint.
- Read your context and instructions
- Compiled the repo atlas
Summary
Generated an evidence-linked documentation atlas for the billing-platform Go monorepo: 7 top-level branches, 31 nodes, dual onboarding tracks, and per-page generation prompts citing real paths.
{
"items": [
{
"title": "Orientation",
"slug": "orientation",
"kind": "section",
"prompt": "Explain what this monorepo is for using README.md and cmd/ entry points; list the 3 files a newcomer should read first.",
"sources": ["README.md", "cmd/", "go.mod"]
},
{
"title": "HTTP Routing",
"slug": "http-routing",
"kind": "page",
"prompt": "Describe how routes are registered and dispatched in the api service, using cmd/api/main.go and internal/transport/http/.",
"sources": ["cmd/api/main.go", "internal/transport/http/"]
},
{
"title": "Beginner Path",
"slug": "beginner-path",
"kind": "guide",
"prompt": "Write a 4-part walkthrough: Go prerequisites vs Python, repo mental model, first feature via tests/, appendices with glossary.",
"sources": ["tests/", "internal/"]
}
]
}
Next steps
- Feed each node's
promptinto your doc generator to produce the wiki pages vendor/was marked third-party and excluded from the concept map- Regenerate after the next major refactor to keep source anchors fresh
Connects securely to your tools. The creator never sees your data.
About this skill
New developers often struggle to navigate complex repositories, wasting hours tracing entry points and guessing at architectural boundaries. Documentation is frequently stale, generic, or missing the "why" behind design choices.
Repo Atlas solves this by analyzing your actual codebase to generate a structured, evidence-linked wiki blueprint. It moves beyond basic README generation to create a layered information map that serves newcomers, reviewers, and senior architects alike.
What it does
- Architectural Mapping identifies subsystems, load-bearing boundaries, and data flow patterns from the directory structure.
- Dual-Path Onboarding generates specific guides for senior engineers needing the big picture and beginners needing a step-by-step path.
- Evidence Linking anchors every documentation topic to concrete file paths and source code locations.
- Prompt Engineering creates specialized instructions for generating detailed page content based on repository evidence.
- Stack Detection automatically identifies frameworks, build systems, and runtimes to tailor the documentation tone.
How it works
- Survey the surface by scanning manifests like
package.json,go.mod, orCargo.tomlalongside the file tree. - Classify the shape of the repository to determine if it is a library, microservice, monorepo, or CLI tool.
- Map the boundaries to locate presentation, application logic, and persistence layers.
- Generate the blueprint as a structured JSON catalogue ready for wiki or documentation site generation.
Frameworks & tools
This skill is language-agnostic and works with any codebase including TypeScript, Python, Go, Rust, Java, and Ruby. It utilizes Mermaid.js for architectural diagrams and GitHub-flavored Markdown for content structure.
Why this beats prompting it yourself
Generic prompts often result in superficial "overview" text that ignores the actual file structure. This skill enforces a strict hierarchy, requires source citations for every claim, and applies senior-level heuristics to identify the real architectural thesis of a project.
Use cases
- Legacy Handover mapping a complex project before a team transition.
- New Hire Onboarding creating a "read this first" path that actually links to the code.
- Wiki Scaffolding building the skeleton for a comprehensive engineering handbook.
- Architecture Audits identifying coupling points and design tradeoffs in a flat codebase.
Known limitations
If only a partial tree is provided, the skill may mark gaps in the documentation. It does not generate the final prose for every page, only the blueprint and specific instructions to do so.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.
Static, structure-based inference only — it maps the repo's shape, not runtime behavior, and page prompts still need a doc generator to produce final wiki content. Very small or completely flat repositories yield deliberately shallow catalogues.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account
Permissions
File Scopes
Read-only repository access plus shell commands to list the file tree and confirm stack signals; no writes and no network required.