Anthropic's announcement of Fable 5.1 and Mythos 5.1 — one base model, two guardrail postures.
- official
- models
- announcement
- fable
12-Factor Agents
HumanLayer's 12 design principles for production-grade LLM agents — the closest thing to a 'best practices' manifesto.
12-Factor Agents is HumanLayer's opinionated manifesto for building production-grade LLM agents. It distils hard-won lessons into 12 principles: own your prompts, own your context window, design for stateless redeploys, and so on. The closest thing the ecosystem has to a 'best practices' document.
What it does
- 12 design principles for production-grade LLM agents from HumanLayer — the closest thing to a best-practices manifesto
- Covers context ownership, tool design, state management, and human escalation
- Written from production agent experience, not theory
When to use it
Read it before shipping your first serious agent — each factor is a checklist item that converts post-launch incidents into pre-launch design decisions.
Getting started
No install — read the principles in the repository. Each factor has its own markdown chapter.
Things to know
Opinionated principles, not a spec — they assume tool-calling agent architectures. Apply them with judgment; not every factor fits every stack.
Repository
humanlayer/12-factor-agentsCharted
Related on the map
Anthropic's canonical essay on agent design — workflows vs agents, and five composable patterns.
- official
- agents
- engineering
- patterns
MCP Registry
GuideOfficial curated registry of MCP servers — discovery, versioning, and metadata for the ecosystem.
- official
- mcp
- registry
- discovery