Architecture Research for Openclaw

An advanced AI agent skill designed to analyze source code, uncover design rationale, and generate professional ELK-based architecture diagrams.

brennerspear
v1.0.0
Feb 25, 2026
0
1k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install architecture-research

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install architecture-research using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Architecture Research?

Architecture Research is a specialized capability within the Openclaw Skills ecosystem that enables AI agents to perform deep-dive technical audits of software systems. Instead of relying on surface-level README files, this skill empowers the agent to parse directory structures, identify entry points, and evaluate dependency injection patterns to build a factual map of a codebase.

Beyond simple code reading, it synthesizes external context such as Architectural Decision Records (ADRs), blog posts, and community discussions to explain the rationale behind technical choices. By integrating with the diagrams skill, it transforms abstract logic into visual ELK JSON graphs and high-quality SVG renders, providing developers with a comprehensive architectural source of truth.

Architecture Research Use Cases

  • Onboarding a new developer to a complex legacy codebase by providing a visual system overview.
  • Documenting data and control flow for critical operations like request handling or deployment pipelines.
  • Evaluating design trade-offs and alternatives during a system refactor or technology migration.
  • Generating automated, up-to-date architecture diagrams for internal documentation repositories.

How Architecture Research Works

  1. The agent clones or browses the target repository to analyze module boundaries, public APIs, and data models.
  2. It conducts a search for external design docs, ADRs, and talks to provide the 'why' behind the implementation.
  3. The system explains the architecture from multiple perspectives, including system overview, data flow, and design decisions.
  4. It generates ELK JSON graphs for each perspective to ensure precise layout control.
  5. The agent utilizes the diagrams skill to render the JSON into SVG files for inline viewing.
  6. A structured research folder is produced, containing the final markdown report, source JSON, and rendered SVGs.

Architecture Research Setup

To use Architecture Research, you must first ensure the diagrams skill is installed as a prerequisite within the Openclaw Skills environment:

clawhub install diagrams

Ensure that elkjs is installed in your local environment to enable the SVG rendering scripts found in the diagrams skill directory.

Architecture Research Data Schema & Taxonomy

The skill produces a structured output directory using the following taxonomy:

File Description
architecture.md The primary deliverable containing prose and embedded SVG diagrams.
prompt.md A record of the original research request and scope.
*.json The ELK JSON source files, allowing for manual edits to diagram layouts.
*.svg Rendered vector graphics embedded into the final report.

Standard reports include a Design Decisions table mapping choices to their rationales and sources.

Architecture Research Advanced Features

  • Multi-angle perspective rendering including layered flow and force-directed peer relationship diagrams.
  • Automated batch processing of multiple ELK JSON files into SVGs or PNGs (macOS).
  • Contextual sourcing that distinguishes between verified documentation and inferred architectural patterns.
  • Fallback support for Mermaid.js or Excalidraw formats based on user preference or tooling constraints.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*