feat: add tested environment episodes and trajectory rewards

[salim4n]

Summary

• Upgrade @ignitionai/agent-trainer-environment from minimal prototype to tested episode runner.
• Add RL bridge from EpisodeResult to Trajectory, deterministic trajectory reports, and offline policy evaluation compatibility.
• Add mocked RAG episode example using search -> rerank -> verify -> answer.
• Update readiness, audit, backlog, milestones, and package docs.

Out of scope

• PPO implementation
• GRPO training
• fine-tuning
• real provider calls
• CLI optimization loop
• IgnitionRAG app integration

Validation

• bun run --filter "./examples/rag-environment-episode" dev
• bun run lint
• bun run typecheck
• bun run test
• bun run build
• bun run pack:check
• git diff --check