References
All source documents, links, and reading material referenced across the engagement.
Claude Agent SDK (workshop, Days 1–3)
- Anthropic — Claude Agent SDK overview
- Anthropic — Building effective agents
- Anthropic — Tool use
- Anthropic — Prompt caching
- Anthropic — Extended thinking
Model Context Protocol (MCP)
Evals
Application fit-gap (Day 4 + Day 5 AM)
- NBS to provide architecture briefs and code/demo access for the ~5 candidate applications at least 1 week before the engagement.
Agent AI guardrails (Day 5 PM)
- Anthropic — Responsible Scaling Policy
- Anthropic — Agent safety best practices
- NIST — AI Risk Management Framework
- OWASP — Top 10 for LLM Applications
MetaWeave reference deployments
- MetaWeave HQ — GitHub
- LifeOSAI — multi-agent operating system (TypeScript monorepo, 27 agents across 6 teams)
- Agent Companies framework —
agentcompanies/v1schema - Ship Management Platform — Ozellar (9 vessels), OneSea (1 vessel)
Engagement artefacts
- Proposal (this site): Proposal summary
- Slide deck (workshop): link TBD
- Lab repos (Days 1–3): links TBD
- Consolidated fit-gap memo (post-engagement): link TBD
- Guardrails framework document (post-engagement): link TBD
Glossary
| Term | Definition |
|---|---|
| Agent | An LLM-driven system that observes, reasons, and acts via tools in a loop until a goal is met or a stop condition fires. |
| Agent SDK | Anthropic’s official SDK for building agents with Claude — handles the loop, tools, and orchestration plumbing. |
| MCP | Model Context Protocol — open standard for connecting agents to external tools, data, and systems. |
| Fit-gap | Structured comparison of an existing system against a reference target (here: modern application architecture). |
| Guardrails | Technical, process, and organisational controls that bound the behaviour of an agent system in production. |
| DX Promotion Group | NBS internal group responsible for digital transformation budget and approvals. |