feat(AIC-3106): add multimodal history support to graph().invoke()

[andrewklatzke]

Summary

• Adds an optional history argument to graph().invoke() (and native graph runners), forwarded to the root node only on the first hop — subsequent nodes keep existing string-based data passing.
• Shared compose_history helpers map LaunchDarkly-canonical multimodal content blocks to each provider’s native format (OpenAI Agents, Claude Agents/Messages, LangChain Agents/Messages).
• Adds a graph-history integration example that asserts the model can identify a generated solid-color PNG from history.

Test plan

• Unit tests for client graph plumbing + all handler/native-graph packages
• Integration: uv run python main.py graph-history travel-agent-flow ""
• Full integration suite (all integration-config.json scenarios, success + failure)
• CI green on this PR

---

Note

Overview
Adds optional history to graph().invoke() and framework native graph runners, passing it only to the root node on the first hop; later nodes keep string-based handoffs.

Introduces shared compose_history and multimodal content helpers in the core client. Provider handlers (Claude/OpenAI/LangChain agents and messages) stop flattening history into the system prompt and instead build structured turns, mapping LD image/text blocks to each SDK’s native shape (streamed envelopes for Claude Agents, merged user turns for Claude Messages, input_image / image_url elsewhere).

Adds a graph-history example (generated red PNG + colour check) and broad unit/integration test coverage for plumbing and wire format.

^{Reviewed by Cursor Bugbot for commit 97dcdcd. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 3 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit da9b323. Configure here.}