Add Future AGI products to Observability / Evaluation / Context Mgmt / Platform Stacks

[SuhaniNagpal7]

This PR adds four Future AGI open-source projects to their most relevant sections in the list.

Added:

• Agent Observability and Telemetry → traceAI. OpenTelemetry-native tracing that auto-instruments 50+ AI frameworks across Python, TypeScript, Java, and C#. Captures prompts, tool calls, retrieval steps, and tokens with no vendor lock-in.
• RAG and Generation Evaluation → ai-evaluation. 50+ metrics, LLM-as-judge, and guardrail scanners (jailbreak, PII, prompt-injection); includes dedicated RAG metrics (context recall, context precision, answer relevancy, faithfulness).
• Context Optimization Engines → agent-opt. Six prompt-optimization algorithms (Random, Bayesian, ProTeGi, Meta-Prompt, PromptWizard, GEPA) for evaluation-driven tuning against any LiteLLM-supported model.
• Platform Stacks and Hosted Agent Runtimes → Future AGI. Open-source self-hostable end-to-end agent engineering and optimization platform unifying tracing, evaluation, simulation, datasets, gateway, and guardrails.

All entries placed in the production / observability subsections matching the existing tooling pattern (Entroly, LangSmith, OpenTelemetry semconv, Puppyone). No paper-style entries.