kitaru

Guide for writing Kitaru durable workflows and operational control paths. Use when creating or refactoring Kitaru flows, checkpoints, waits, logging, artifacts, tracked LLM calls, replay/resume/retry flows, KitaruClient usage, CLI commands, MCP operations, deployments, secrets, or adapter integrations for PydanticAI, OpenAI Agents, LangGraph, Claude Agent SDK, and Gemini Interactions. Triggers on mentions of kitaru, @flow, @checkpoint, kitaru.wait, kitaru.log, kitaru.save, kitaru.load, KitaruClient, replay, resume, retry, `kitaru executions ...`, MCP tools, `KitaruAgent`, `KitaruRunner`, `KitaruGraphRunner`, `KitaruClaudeRunner`, `KitaruGeminiInteractionsRunner`, `GeminiInteractionRequest`, `wait_for_input`, `wait_for_approval`, `wait_for_interrupt`, `requires_action`, Antigravity, or migration from deprecated `wrap(...)`.

Migrate existing Claude Agent SDK code to Kitaru's Claude Agent SDK adapter. Use when code mentions claude_agent_sdk, claude_agent_sdk.query, query, ClaudeAgentOptions, ClaudeSDKClient, ResultMessage, SystemMessage, session_id, resume, cwd, allowed_tools, Bash, Read, Write, Edit, MCP, hooks, permissions, file checkpointing, transcript, workspace, KitaruClaudeRunner, ClaudeRunRequest, ClaudeRunResult, ClaudeCapturePolicy, checkpoint_strategy, or invocation. Helps classify direct, approximate, and absent mappings, keep Claude-owned internals inside one invocation boundary, and produce a migration report with replay, workspace, session, tool, Bash, MCP, hook, and privacy risks called out.

Migrate existing Gemini Interactions, Google GenAI, and Antigravity managed agent code to Kitaru's Gemini Interactions adapter. Use when code mentions google.genai, genai.Client, client.interactions.create, client.interactions.get, Interaction, interaction_id, previous_interaction_id, requires_action, function_call, function_result, background, store, poll, steps, output_text, Antigravity, managed agents, environment, Vertex, API key, KitaruGeminiInteractionsRunner, GeminiInteractionRequest, GeminiInteractionResult, GeminiInteractionCapturePolicy, or cache_identity. Helps classify direct, approximate, and absent mappings, preserve Google-owned hosted interaction internals, handle requires_action and polling safely, and produce a migration report.

Migrate existing LangGraph and LangChain agent code to Kitaru's LangGraph adapter. Use when code mentions LangGraph, StateGraph, CompiledStateGraph, graph.invoke, graph.ainvoke, stream, astream, interrupt, Command, thread_id, checkpointer, InMemorySaver, MemorySaver, create_agent, KitaruGraphRunner, LangGraphRunRequest, LangGraphRunResult, KitaruLangGraphMiddleware, build_resume_request, wait_for_interrupt, checkpoint_strategy, graph_call, calls, callbacks, event streams, Deep Agents, or checkpointers. Helps classify direct, approximate, and absent mappings, choose graph_call vs middleware-backed calls boundaries, preserve LangGraph ownership of graph state, and produce a migration report.

Migrate existing OpenAI Agents SDK code to Kitaru's OpenAI Agents adapter. Use when code mentions agents.Agent, Runner.run, Runner.run_sync, KitaruRunner, OpenAIRunRequest, OpenAIRunResult, wait_for_approval, build_resume_request, function_tool, RunConfig, handoffs, context, approval/resume flows, or checkpoint_strategy. Helps classify direct, approximate, and absent mappings, choose runner_call vs calls boundaries, preserve OpenAI Agents SDK ownership of the agent turn, and produce a MIGRATION_REPORT.md with replay, context, approval, tool, and side-effect risks called out.