Automatically send Claude Code conversation traces to LangSmith for monitoring and analysis. Prerequisites: jq (brew install jq on macOS or sudo apt-get install jq on Linux), curl and uuidgen (usually pre-installed), LangSmith account and API key. Configuration: Install langsmith-tracing setting (npx claude-code-templates@latest --setting telemetry/langsmith-tracing) or manually add to .claude/settings.local.json the following environment variables: TRACE_TO_LANGSMITH=true, CC_LANGSMITH_API_KEY=lsv2_pt_..., CC_LANGSMITH_PROJECT=project-name, CC_LANGSMITH_DEBUG=true (optional). How it works: Runs in background on Stop event after each Claude response, reads conversation transcript, converts to LangSmith format, sends to LangSmith API, groups by thread_id for session continuity. Debugging: Check logs at ~/.claude/state/hook.log. Privacy note: System prompts not included in traces.
References specific files for sparse matrix operations, explains transform usage, compares with previous approaches, and sets data formatting context for development.
Sets context for model training by referencing pipeline files, establishing working context, and preparing for pipeline work with relevant documentation.
Loads LLM configuration files to context, importing specific terminology, model configurations, and establishing baseline terminology for AI discussions.
Heterogeneous team via the multiagent coordinator config — a coordinator runs three specialists (web-search researcher, file-reading librarian, rules-based pricer) with scoped toolsets to assemble a sales proposal. Covers the multiagent field, the thread_created / thread_message_received event types, and per-role to…
Entry-point tutorial for the Claude Managed Agents API. Walks through agent / environment / session creation, file mounts, and the streaming event loop by getting an agent to fix three planted bugs in a calc.py package.