Semantic code knowledge graph builder. Scans repos, builds structured feature trees and spec graphs, detects code-spec drift, generates documentation. Six skills: build, check, update, render, preview, help. Plus /shadowrepo-review pre-commit alignment check and /shadowrepo-pr-comment GitHub-ops draft-PR wrap.
Show all ShadowRepo capabilities and usage. Use when: "shadowrepo help", "what can shadowrepo do", "how to use shadowrepo".
Open a draft GitHub PR and post a ShadowRepo-aware summary as its first comment. Reads .shadowrepo/features.json + coverage.json + specs.json, renders a markdown context block, wraps `gh pr create --draft` then `gh pr comment`. Degrades to summary-only when `gh` is missing or unauthenticated. Use when: "/shadowrepo-pr-comment", "open PR with shadowrepo summary", "draft pr with context", "post shadowrepo summary on pr".
Cold-start a ShadowRepo for a codebase. Scans the entire repo, builds a feature tree and spec graph, writes structured JSON to .shadowrepo/. Use when: "build shadowrepo", "scan this project", "create a shadow repo", or first-time setup.
Detect drift between code and existing ShadowRepo specs. Compares git changes against spec anchors to find content changes, orphaned anchors, and uncovered new files. Use when: "check for drift", "is shadowrepo up to date", "what specs are stale".
Impact assessment: given a feature doc or PRD, analyze how it would affect the existing ShadowRepo. Runs check first to ensure specs are current. v0.2 — stub implementation. Use when: "what would this change affect", "preview impact", "impact assessment".
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
The context layer for codebases — for both coding agents and humans.
A semantic knowledge graph of what your code does and why, kept in sync by an AI agent.
Install · What's in the box · How it works · Build from source
The decisions behind code — why JWT over sessions? why this schema? why five pipeline stages? — are scattered across PR threads, commit messages, and tribal knowledge. AI agents can read syntax, but humans can't see the reasoning.
ShadowRepo scans your repo, extracts structured semantic specs into a .shadowrepo/ directory, organizes them into a feature tree, and detects when code drifts from those specs. Then it surfaces all of that in your editor.
ShadowRepo ships as three connected pieces that can be installed independently.
Install from the VS Code Marketplace or Open VSX. Or in VS Code: Extensions → search "ShadowRepo".
The sidebar reads .shadowrepo/*.json from your workspace and renders a navigable feature tree with click-to-jump anchors.
.shadowrepo/In Claude Code:
/plugin marketplace add waynewangyuxuan/shadowrepo-extension-package
/plugin install shadowrepo
This gives you slash commands like /shadowrepo-build, /shadowrepo-check, /shadowrepo-update, /shadowrepo-review, and /shadowrepo-pr-comment.
The Claude plugin pulls this in automatically via npx. To use it from another MCP-compatible client:
npx -y shadowrepo-mcp-server@latest
Or pin it in your client's MCP config:
{
"mcpServers": {
"shadowrepo": {
"command": "npx",
"args": ["-y", "shadowrepo-mcp-server@latest"]
}
}
}
| Package | Description | Distribution |
|---|---|---|
packages/vscode-extension | Sidebar tree of features + specs + anchors. Spec dossier webviews. Auto-refresh on .shadowrepo change. | VS Code Marketplace, Open VSX |
packages/claude-plugin | The agent that scans code and produces/maintains .shadowrepo/. Six core skills + two slash commands. | Claude Code plugin marketplace |
packages/mcp-server | MCP server exposing incremental .shadowrepo/ read/write tools. | npm: shadowrepo-mcp-server |
packages/shared | Shared TypeScript types for the .shadowrepo/*.json schema. | Internal, bundled into the above. |
┌─────────────────┐ scans ┌──────────────┐ reads ┌──────────────────┐
│ Claude Plugin │ ─────────► │ .shadowrepo/ │ ─────────► │ VS Code Sidebar │
│ (the agent) │ │ *.json │ │ (humans) │
└─────────────────┘ └──────────────┘ └──────────────────┘
│ ▲
│ incremental updates │
└──────► MCP Server ───────────┘
(shadowrepo-mcp-server)
/shadowrepo-build once. The plugin walks your repo and writes a feature tree + spec graph to .shadowrepo/./shadowrepo-update (or live MCP calls) keep the graph in sync. /shadowrepo-check flags drift.See DEMO.md for a step-by-step walkthrough (clone, install, build, sideload).
MIT. See LICENSE.
npx claudepluginhub waynewangyuxuan/shadowrepo-extension-package --plugin shadowrepoAutonomous multi-agent software development workflow: Intake → Planning → Scheduler → Executor × N → Review. Takes a one-shot human brief and delivers a software project end-to-end as a batch job. Do NOT use for interactive pair-programming.
AI tests skills, humans write them. Stress-test AI agent skills across diverse scenarios, get diagnostic evidence and improvement proposals.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Connect to Atlassian products including Jira and Confluence. Search and create issues, access documentation, manage sprints, and integrate your development workflow with Atlassian's collaboration tools.
Streamline engineering workflows — standups, code review, architecture decisions, incident response, and technical documentation. Works with your existing tools or standalone.
Optimize business operations — vendor management, process documentation, change management, capacity planning, and compliance tracking. Keep your organization running efficiently.
Access official Microsoft documentation, API references, and code samples for Azure, .NET, Windows, and more.
Make your AI agent code with your project's architecture, rules, and decisions.