Toolkit

Interactive tools, research references, and guides for building with AI. Grounded in peer-reviewed research and practical enterprise experience.

New

RossLabs Claude Code Plugins

16 open-source plugins for Claude Code — build-loop, ibr, navgator, bookmark, claude-code-debugger, and more. Version-tracked, installable via Claude Code's standard plugin workflow.

# Add the marketplace — use owner/repo format (not the GitHub web URL)
claude plugin marketplace add tyroneross/RossLabs-AI-Toolkit

# Install any plugin from it
claude plugin install build-loop@rosslabs-ai-toolkit
claude plugin install ibr@rosslabs-ai-toolkit
claude plugin install navgator@rosslabs-ai-toolkit

Browse all 16 plugins Each card links to its GitHub repo. Plugins with a writeup also link into Projects.

build-loop

v0.3.0

developer tools

Orchestrated 5-phase development loop — Assess, Plan, Execute, Review, Iterate — plus optional Learn. Opus 4.7 orchestrator with NavGator / debugger / tracer bridges.

build-loop orchestration agents

ibr

v1.0.1

developer tools

UI validation — verify implementations match intent with live page scanning and visual regression.

ui validation visual-regression

navgator

v0.6.2

developer tools

Architecture tracking — map dependencies, analyze impact, and visualize your stack before you change it.

architecture dependencies impact-analysis

bookmark

v0.3.2

developer tools

Session context continuity — auto-save and restore context across compactions and terminal closures.

context snapshots session-continuity

claude-code-debugger

v1.8.1

developer tools

Debugging memory — learn from past bugs with verdict-based retrieval and pattern extraction.

debugging memory incident-tracking

research

v0.3.2

developer tools

Token-efficient research knowledge base — SQLite FTS5, source tier scoring, claim verification, lifecycle management.

research knowledge-base fact-checking

agent-builder

v0.1.0

developer tools

Design and evaluate agentic harnesses — methodology playbooks plus an empirical catalog of architectures, memory substrates, and production patterns.

agent harness architecture

prompt-builder

v0.1.1

developer tools

Prompt Policy Engine — classify, diagnose, rewrite, and score prompts by model tier and deployment.

prompt policy scoring

pyramid-principle

v0.1.2

writing

Barbara Minto's Pyramid Principle as composable writing skills — short-form, long-form, presentations, and audit.

writing minto communication

showcase

v0.1.1

developer tools

Dev asset capture — screenshots and video clips during development for blog and website content.

screenshots video dev-assets

spectra

v0.1.1

developer tools

Content capture for marketing — screenshots, videos, and usage sequences across web, macOS, iOS, and watchOS.

video-capture marketing demo-recording

mockup-gallery

v0.4.1

design

Visual mockup review with component-level ratings, auto-save to file, and Claude Code integration.

design-review mockups ratings

replit-migrate

v0.1.1

developer tools

Migrate Replit apps to web (Vercel) or native (iOS/macOS) with encoded lessons from real migrations.

replit migration ios

scraper-app

v0.5.0

developer tools

Intelligent web scraper for extracting blog and news content from any website.

scraper blog rss

stratagem

v0.1.0

research

Market research agent with document processing, web scraping, SEC filings, and financial analysis.

research sec financial-analysis

agent-astronomer

v0.1.0

developer tools

Query your local Claude Code skill, agent, and plugin library from any conversation. Wraps the Agent Astronomer CLI as MCP tools.

skills plugins catalog

Tools & Artifacts

Prompt Decision Aid

14 research-backed prompting techniques. Filter by use case, compare approaches, and find the right method.

Interactive matcher for chain-of-thought, few-shot, skeleton-of-thought, and more.

FlowDoro

Floating pomodoro timer with flow-state tracking and break nudges.

25/5 cycles that surface as a draggable popup you can position anywhere on screen.

Agent Architectures

Five architecture types, seven frameworks, seven coordination patterns — the reference corpus for building agents in 2026.

Interactive taxonomy of how agents are actually built today, grounded in 368 sources from Anthropic, Deloitte, Gartner, and Stanford.