codex/increment-3521
6 次代码提交
| 作者 | SHA1 | 备注 | 提交日期 | |
|---|---|---|---|---|
 Asim Aslam
|
4a09b4469a |
Add flow execution timeout option (#3164)
Co-authored-by: Codex <codex@openai.com> |
||
|
Asim Aslam
|
9d39ed3292 |
Add OpenTelemetry spans for stepped flows (#3098)
Co-authored-by: Codex <codex@openai.com> |
||
|
Asim Aslam
|
105cc9d2fb | Add flow retry backoff (#3053) | ||
|
Asim Aslam
|
9fdcc24cce |
Implement durable execution and scoped state management for flows (#2972)
* docs: design note for flow steps + Checkpoint durable execution
* docs: fold in durable-execution decisions (State struct, single Step, run retention, retry)
* docs: rename State.Payload to State.Data
* feat(flow): ordered steps + Checkpoint durable execution
A flow can now be an ordered list of steps (a task with stages) instead
of a single LLM turn. State carries typed Data plus a Stage marker; each
step is checkpointed before and after via a pluggable Checkpoint
(store-backed by default), so a run survives a crash and resumes where it
stopped without re-running completed steps. Flow-level Retry with a
per-step override; runs retained for audit unless DeleteOnSuccess.
Step actions: Call (RPC), LLM (augmented turn), Dispatch (to an agent),
or any StepFunc. Single-step and agent-dispatch flows are unchanged.
* feat(flow): top-level re-exports + durable flow example
Expose the step/checkpoint API from the micro package (FlowSteps,
FlowStep, FlowState, FlowRetry, FlowWithCheckpoint, FlowCall/LLM/Dispatch,
Checkpoint, StoreCheckpoint) and add a runnable, key-free example
demonstrating crash + resume.
* docs: document durable flow steps (guide, README, CLI help)
* docs: blog post + changelog for durable workflows
* fix(flow): scope checkpoint keys by flow name (flow/{name}/runs/{id})
Run keys were flow/runs/{id} — a single global keyspace shared by every
flow on the default store. Namespace them by flow name so each flow's
state is kept apart. StoreCheckpoint now takes a scope argument (the flow
passes its name by default).
* feat(store): Scope handle; scope agent and flow state by name
Add store.Scope(s, database, table) — a store handle that confines every
operation to a database/table without mutating the shared store, so
co-located components don't clobber each other's table (the failure mode
of the global Init(Table(...)) approach).
Use it to keep each agent's memory and plan in its own table
(agent/{name}) and each flow's runs in its own (flow/{name}), instead of
one global table partitioned only by key prefix. Services already scope
by service name.
* feat: consistent state model — service store scoping, flow registry, list/history CLI
- service: scope store via store.Scope (database service / table name),
retiring the Init(store.Table(name)) global-mutation hack; bridge the
default store so handlers using store.DefaultStore stay isolated.
- flow: register in the registry as type=flow while running (with trigger
and step count), deregister on Stop. Live discovery, like agents.
- cli: micro flow list (registry), micro flow runs <name> (durable store),
micro agent history <name> (durable store). list = running, runs/history
= durable, mirroring the service model.
* test: mini-universe end-to-end harness + scheduled GitHub Action
internal/harness/universe boots a small but real go-micro world — four
services, a durable checkout flow that crashes at payment and resumes,
and a guardrailed agent with a tool wrapper reached over RPC — drives the
scenario, asserts the end state (10 checks), and shuts down. Everything
is real except the LLM (mocked), so it's deterministic and needs no key;
-provider anthropic runs it live. Exits non-zero on failure, so it's an
end-to-end test, not just a demo.
Adds .github/workflows/universe.yml (push/PR/daily/dispatch) running the
universe + existing harnesses on the mock provider, plus an opt-in job
that runs live when ANTHROPIC_API_KEY is set. 'make harness' runs them
locally.
* ci: run the live universe job against AtlasCloud (ATLASCLOUD_API_KEY)
* ci: run the live universe job only on schedule or manual dispatch
The deterministic mock job still runs on push/PR/daily; the live
(AtlasCloud) job runs daily and on manual workflow_dispatch only, so
changes don't burn API credits on every PR but can still be checked
against a real model on demand.
---------
Co-authored-by: Claude <noreply@anthropic.com>
|
||
|
Asim Aslam
|
e416ea4a75 |
Enhance agent workflows with guardrails and documentation updates (#2952)
* docs: map go-micro onto Anthropic's workflows-vs-agents taxonomy - new guide 'Agents and Workflows': adopts Anthropic's Building Effective Agents vocabulary — workflow (predefined path) = flow, agent (dynamic self-direction) = agent — maps the augmented-LLM building block and the five workflow patterns onto go-micro, and shows routing (chat router) and orchestrator-workers (conductor + plan/delegate) are already native. - flow package doc reframed as a workflow (predefined path) per the same taxonomy, with guidance on flow vs agent. - nav + README link the new guide. * feat: agent guardrails — step limit and tool approval hook Anthropic's Building Effective Agents stresses stopping conditions and human-in-the-loop checkpoints for autonomous agents. Add both as plain options enforced at the tool-handler choke point — no provider changes, no new abstraction: - MaxSteps(n): bound tool executions per Ask; beyond the limit, actions are refused and the model is told to stop and summarize. - ApproveTool(fn): gate each action before it runs; returning false blocks it and surfaces the reason to the model. The internal plan tool is never gated. Exposed at the micro package (AgentMaxSteps, AgentApproveTool, ApproveFunc). Tests cover the limit, blocking, and that plan is not gated. Guardrails section of the agents-and-workflows guide updated from 'active work' to documented options. * feat: flow can dispatch to an agent (flow triggers, agent reasons) Unify the engine without collapsing the workflow/agent distinction. A Flow with Agent set hands each event's rendered prompt to a named registered agent over RPC (Agent.Chat) instead of running its own LLM step — so the workflow stays the deterministic trigger and the agent is the reasoning engine, with its plan, delegate, memory, and guardrails. A plain flow is unchanged (single augmented-LLM step). - flow.Agent(name) / micro.FlowAgent(name); flow stores the client and skips model setup when dispatching. - test: dispatch routes to comms.Agent.Chat with the rendered prompt and records the reply. - guide: 'Flow triggers, Agent reasons' section. --------- Co-authored-by: Claude <noreply@anthropic.com> |
||
|
Asim Aslam
|
1bc886fa82 |
Introduce Agent abstraction and integrate with chat router (#2939)
* docs: Agent interface design sketch Proposes Agent as a top-level abstraction alongside Service in the micro package. Agent manages services — scoped tools, system prompt, conversation memory, registry-discoverable. Design only, no implementation. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: Agent as a first-class abstraction Introduce micro.NewAgent() alongside micro.New() — Agent is to intelligence what Service is to capability. Agent interface: - Chat(ctx, message) (*Response, error) — core interaction method - Run() — registers in registry, subscribes to broker, blocks - Stop() — graceful shutdown - Scoped tools — only sees endpoints of its assigned services - Persistent memory — conversation history stored in store - Agent-to-agent — communication via broker topics Top-level API: agent := micro.NewAgent("task-mgr", micro.AgentServices("task"), micro.AgentPrompt("You manage tasks."), micro.AgentProvider("anthropic"), ) agent.Run() https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: wire agents into chat router, add micro agent CLI, expose Flow Three top-level abstractions: micro.New("task") — Service (capability) micro.NewAgent("task-mgr") — Agent (intelligence) micro.NewFlow("onboard-user") — Flow (event-driven orchestration) micro chat as router: - Discovers agents from registry on startup - Single agent: routes directly - Multiple agents: LLM classifies intent, dispatches to right agent via route_to_agent tool - No agents: falls back to current direct-service behaviour - Banner shows discovered agents micro agent CLI: - micro agent list — shows registered agents and their services - micro agent describe <name> — shows agent details from registry https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: move flow to top level, update docs for three abstractions Package structure now consistent: service/ — Service (capability) agent/ — Agent (intelligence) flow/ — Flow (event-driven orchestration) ai/flow/ kept as backward-compatible re-export. Updated across all surfaces: - CLAUDE.md: added agent/ and flow/ to project structure - README: added "Building Agents" section with NewAgent() examples, updated features table (Agents, Flows, Chat router), CLI table (agent list, agent describe), docs links - Website: features grid shows Services, Agents, Flows as the three pillars alongside generation, MCP, and pluggable architecture - micro.go: Flow imported from top-level flow/ package https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * docs: rewrite getting-started, fix ai-integration import paths Getting started now covers all three abstractions: - Service (write handlers, micro run, templates) - Agent (micro.NewAgent, scoped tools, memory, CLI) - Flow (event-driven LLM orchestration) Leads with prompt-based generation, then manual service creation. ai-integration.md: fixed flow import path from go-micro.dev/v5/ai/flow to go-micro.dev/v5/flow, updated stack diagram to show agent/flow/chat. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * blog: Introducing micro.NewAgent() Post 16 — announces Agent as a first-class abstraction. Shows the API (NewAgent, AgentServices, AgentPrompt, AgentProvider), scoped tools, persistent memory, multi-service agents, multi-agent systems, and the three-abstraction comparison table (Service/Agent/Flow). https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: fix agent registration, blog post 16 Agent registration: - Add node with address so mDNS can discover agents - Store type and services in node metadata (mDNS requirement) - Connect broker before subscribing, non-fatal if broker unavailable - Print registration confirmation on Run() Agent/chat discovery: - Check both service-level and node-level metadata for type=agent (mDNS stores metadata on nodes, not services) Blog post 16: "Introducing micro.NewAgent()" — announces the Agent abstraction with code examples, comparison table, multi-agent patterns. Tested end-to-end: micro run → micro agent list discovers the agent → micro chat routes to it → agent calls service endpoints. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: agents are proper services with RPC Chat endpoint Refactored agent to use server.Server instead of fake registry entries. An agent now: - Creates a real RPC server with server.Name(agentName) - Registers an Agent.Chat handler callable via standard RPC - Sets server metadata type=agent, services=x,y for discovery - No more fake addresses or broker hacks micro chat calls agents via RPC (client.Call) instead of creating local agent instances. The registry stays clean — agents are real services with real endpoints. Removed broker dependency from agent options. Agent-to-agent communication is just RPC like everything else. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: agent uses proto-defined RPC interface Added agent/proto/agent.proto with Agent service definition: rpc Chat(ChatRequest) returns (ChatResponse) Agent now implements the generated AgentHandler interface and registers via pb.RegisterAgentHandler. The Chat endpoint is a standard proto-based RPC callable by any go-micro client. Renamed the programmatic API from Chat() to Ask() to avoid collision with the proto handler method name. micro chat calls agents via standard RPC with JSON-encoded request/response — no special types needed on the caller side. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd * feat: generate agent alongside services, update all docs micro run --prompt now generates an agent binary that manages all the generated services. The agent reads MICRO_AI_PROVIDER and MICRO_AI_API_KEY from the environment. micro run propagates these when started with --prompt. Run banner shows services and agents separately. Updated README, getting-started guide, and landing page to show the complete flow: generate → services + agent start → micro chat routes to agent → agent orchestrates services. https://claude.ai/code/session_01QTp4SshuVmLAvvEGJe4TJd --------- Co-authored-by: Claude <noreply@anthropic.com> |