Changelog

All notable changes to the Open AI Engineering Standard are documented here.

Format: [Version] — YYYY-MM-DD Each entry includes: Added, Changed, Deprecated, Removed, Fixed, Security.

---

[1.0.0] — 2026-07-16

Added

• Complete 10-level AI Engineering Maturity Model (L0–L10)
• 12 AI Engineering Patterns with full 13-component specification
• 18-stack AI Cookbook
• 26 reusable agent skills
• 20 production agent definitions
• 19 subagent definitions
• 12 lifecycle hooks
• 24 MCP integration specifications
• Complete AI SDLC with 26 stages
• 15 production-ready SDLC prompts
• Enterprise security, governance, and compliance guides
• Anti-pattern registry across all domains
• Evaluation framework setup guides (RAGAS, DeepEval, Promptfoo)
• GitHub automation templates
• Curated papers and resources

Standard Decisions (v1.0)

• Prompt format: XML-structured (system → user → output_format → constraints)
• Context placement: Critical information at top and bottom (anti-"lost in the middle")
• Agent harness: Provider-neutral, deterministic wrapper required
• Evaluation: DeepEval for CI gates, RAGAS for RAG pipelines
• Observability: OpenTelemetry-native with LangSmith/Langfuse integration
• Prompt management: Versioned artifacts (Langfuse or Braintrust)
• Security posture: Least privilege by default, OAuth 2.1 for MCP
• Cost control: Model routing enforced, token budgets mandatory

---

Versioning Roadmap

Version	Target	Focus
v1.0	Q3 2026	Foundations + Prompt + Context + Skill Engineering
v2.0	Q4 2026	Agent + Multi-Agent Engineering
v3.0	Q1 2027	LLMOps + AI SDLC
v4.0	Q2 2027	Enterprise AI + AI Org Playbook
v5.0	Q3 2027	Complete Patterns Library
v6.0	Q4 2027	Complete Cookbook
v7.0	Q1 2028	Memory + Knowledge Graphs deep dive
v8.0	Q2 2028	Advanced Evaluation Frameworks
v9.0	Q3 2028	AI Governance Standards
v10.0	Q4 2028	Full living standard — all 10 levels complete