Engineering Blueprint
Autonomously Conduct AI Research with Expert Skills
Comprehensive library of 98 production-ready skills enabling AI agents to autonomously manage the full AI research lifecycle from ideation to paper publication.
Install Command
Run this command to deploy the blueprint to your environment.
What does this do
Modern AI researchers waste time debugging infrastructure and learning dozens of specialized tools instead of testing hypotheses and making discoveries. Building autonomous AI agents requires integrating expertise across 23 distinct domains—from model architecture to paper writing—without a unified system. Teams need a centralized, battle-tested skill library that covers the complete research workflow.
How It Works
You install 98 modular AI research skills organized across 8 functional domains. Each skill provides 300KB+ of production-ready guidance with real code examples, troubleshooting guides, and official documentation. Use them individually for specific tasks (fine-tune a model, set up inference, write papers) or activate the autoresearch orchestrator to let AI agents autonomously conduct end-to-end research cycles, routing through domain skills as needed.
Key Features
98 Production Skills
Comprehensive coverage of the entire AI research lifecycle: model training, optimization, inference, evaluation, safety, applications, and more—all battle-tested with real code examples
Unified Orchestration
Autoresearch skill uses two-loop architecture to autonomously route research tasks across domain skills, managing the full workflow from idea to paper
300KB+ Documentation
Each skill includes 300KB+ sourced from official repos, GitHub issues, and production workflows—never vendor marketing, always battle-tested guidance
Multi-Agent Support
One-command installation across 8+ AI coding agents: Claude Code, OpenCode, Cursor, Codex, Gemini CLI, and more with automatic detection
Research Artifacts
Convert research outputs into Agent-Native Research Artifacts with falsifiable claims, exploration graphs, and grounded evidence for auditability and reproducibility
Key Features
98 Production Skills
Comprehensive coverage of the entire AI research lifecycle: model training, optimization, inference, evaluation, safety, applications, and more—all battle-tested with real code examples
Unified Orchestration
Autoresearch skill uses two-loop architecture to autonomously route research tasks across domain skills, managing the full workflow from idea to paper
300KB+ Documentation
Each skill includes 300KB+ sourced from official repos, GitHub issues, and production workflows—never vendor marketing, always battle-tested guidance
Multi-Agent Support
One-command installation across 8+ AI coding agents: Claude Code, OpenCode, Cursor, Codex, Gemini CLI, and more with automatic detection
Research Artifacts
Convert research outputs into Agent-Native Research Artifacts with falsifiable claims, exploration graphs, and grounded evidence for auditability and reproducibility
Tools in this Blueprint
About This Blueprint
- License
- MIT
- Industry
- Technology
- Skills
- 1 workflows, 0 sub-skills, 97 standalone
You might also like
AI + Monitoring what AI influencers and... Playbook
Stay ahead of AI developments with automated daily digests of what key thought leaders are discussing—surfacing emerging tools and model changes before they impact your roadmap.
Ben D.
Sr Dir, AI Engineering
AI + B2B marketing teams produce generic,... Playbook
Transform generic B2B marketing copy into revenue-focused messages that emphasize commercial outcomes. This embedded copywriter enforces discipline across email, landing pages, and thought leadership
Danielle M.
Marketing specialist
G2 Review Intelligence to SC Discovery Brief
Equip Solutions Consultants with data-backed discovery questions and post-call validation by matching prospect pain points to trending G2 reviews—turn generic call prep into targeted, insight-driven c
Matt R.
VP, Solutions and Success