Ts langchain

[Adarsh9977]

Summary

Added langchain TS sdk

Test plan

• cd sdk/typescript npm test -w @secureagentics/adrian-langchain.

Checklist

• CLA signed (see CLA.md)
• Tests pass locally
• Docs updated where needed
• British English; no em-dashes; no marketing fluff

[yanny-sec]

Thank you for another integration! 👍

Two things from my testing so far:

• deriveAgentId() in core/src/identity.ts justs return "default" for everything, so every event gets the same agent id and the AgentContextTracker parent/hierarchy logic can never actually fire. Python derives a real id from the LangGraph checkpoint_ns, and that's what powers all the S1–S8 scenarios (supervisor / handoff / hierarchy / swarm / etc.)
• Related: invocation_id is inherit-only and nothing in the SDK ever creates one and there's no Pregel/graph-entry hook, so a multi-agent run doesn't get auto-grouped (everything falls back to "no_invocation" unless the app calls runWithInvocationId itself). On the Python side that id also scopes the classifier's history window per (session, invocation, agent_id), so it matters for context-aware detection, not just tracing.

If you really want to stress test a new integration you should try multi agent scenarios to ensure that all the chat history and tool calls are correctly placed in each sliding window.