{"id":35127433,"url":"https://github.com/agent-field/agentfield","last_synced_at":"2026-06-15T05:01:39.132Z","repository":{"id":323557210,"uuid":"1089925776","full_name":"Agent-Field/agentfield","owner":"Agent-Field","description":"Build, run and scale AI agents like API and microservices - observable,auditable and identity-aware from day one.","archived":false,"fork":false,"pushed_at":"2026-05-28T22:34:22.000Z","size":268756,"stargazers_count":2014,"open_issues_count":85,"forks_count":321,"subscribers_count":21,"default_branch":"main","last_synced_at":"2026-05-29T00:16:43.458Z","etag":null,"topics":["agent","agent-auth","agent-authentication","agent-indentity","agent-scaling","agentic-ai","ai","ai-backend","aiagent","anthropic","cloud-native","genai","go","kubernetes","llm","multiagent","multiagent-systems","python","rag","typescript"],"latest_commit_sha":null,"homepage":"http://www.agentfield.ai","language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Agent-Field.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":"docs/CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":"CODEOWNERS","security":"SECURITY.md","support":"SUPPORT.md","governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2025-11-05T02:04:44.000Z","updated_at":"2026-05-28T22:40:36.000Z","dependencies_parsed_at":"2026-03-12T22:02:35.340Z","dependency_job_id":null,"html_url":"https://github.com/Agent-Field/agentfield","commit_stats":null,"previous_names":["agent-field/agentfield"],"tags_count":322,"template":false,"template_full_name":null,"purl":"pkg:github/Agent-Field/agentfield","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Agent-Field%2Fagentfield","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Agent-Field%2Fagentfield/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Agent-Field%2Fagentfield/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Agent-Field%2Fagentfield/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Agent-Field","download_url":"https://codeload.github.com/Agent-Field/agentfield/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Agent-Field%2Fagentfield/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33893323,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-04T02:00:06.755Z","response_time":64,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent","agent-auth","agent-authentication","agent-indentity","agent-scaling","agentic-ai","ai","ai-backend","aiagent","anthropic","cloud-native","genai","go","kubernetes","llm","multiagent","multiagent-systems","python","rag","typescript"],"created_at":"2025-12-28T03:33:20.285Z","updated_at":"2026-06-04T07:00:43.170Z","avatar_url":"https://github.com/Agent-Field.png","language":"Go","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\n# AgentField — The AI Backend\n\n### **Build and scale AI agents like APIs. Deploy, observe, and prove.**\n\n*AI has outgrown chatbots and prompt orchestrators. Backend agents need backend infrastructure.*\n\n[![Stars](https://img.shields.io/github/stars/Agent-Field/agentfield?style=flat\u0026logo=github\u0026logoColor=e8e5dc\u0026color=0c0b09\u0026labelColor=8b7355)](https://github.com/Agent-Field/agentfield/stargazers)\n[![License](https://img.shields.io/badge/license-Apache%202.0-0c0b09.svg?style=flat\u0026labelColor=8b7355)](LICENSE)\n[![Downloads](https://img.shields.io/endpoint?url=https%3A%2F%2Fgist.githubusercontent.com%2Fsantoshkumarradha%2Fd98e2ad73502b4075f6a5f0ae4f5cae5%2Fraw%2Fbadge.json\u0026style=flat\u0026logo=download\u0026logoColor=e8e5dc\u0026color=0c0b09\u0026labelColor=8b7355\u0026cacheSeconds=3600)](https://github.com/Agent-Field/agentfield)\n[![Coverage](https://img.shields.io/endpoint?url=https%3A%2F%2Fgist.githubusercontent.com%2Fsantoshkumarradha%2F433fb09c2d54c3c2589125cfd3eb14a2%2Fraw%2Fbadge.json\u0026style=flat\u0026logo=github\u0026logoColor=e8e5dc\u0026color=0c0b09\u0026labelColor=8b7355\u0026cacheSeconds=3600)](https://github.com/Agent-Field/agentfield/actions/workflows/coverage.yml)\n[![Last Commit](https://img.shields.io/github/last-commit/Agent-Field/agentfield?style=flat\u0026logo=git\u0026logoColor=e8e5dc\u0026color=0c0b09\u0026labelColor=8b7355)](https://github.com/Agent-Field/agentfield/commits/main)\n[![Discord](https://img.shields.io/badge/discord-join%20us-0c0b09.svg?style=flat\u0026labelColor=8b7355\u0026logo=discord\u0026logoColor=e8e5dc)](https://discord.gg/aBHaXMkpqh)\n\n\u003c!-- All outbound agentfield.ai links must use UTM parameters, not direct URLs. See assets/utm-links.csv for the full list — target URLs there are for verification and reference only, always use the UTM version in the README. Update both the README and the CSV when adding or changing links. --\u003e\n**[Docs](https://agentfield.ai/docs/learn?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-docs)** · **[Quick Start](https://agentfield.ai/docs/learn/quickstart?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-quickstart)** · **[Python SDK](https://agentfield.ai/docs/reference/sdks/python?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-python-sdk)** · **[Go SDK](https://agentfield.ai/docs/reference/sdks/go?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-go-sdk)** · **[TypeScript SDK](https://agentfield.ai/docs/reference/sdks/typescript?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-typescript-sdk)** · **[REST API](https://agentfield.ai/docs/reference/sdks/rest-api?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-rest-api)** · **[Examples](#built-with-agentfield)** · **[Discord](https://discord.gg/aBHaXMkpqh)**\n\n\u003c/div\u003e\n\n\u003cdiv align=\"center\"\u003e\n\u003ca href=\"https://agentfield.ai/docs/build/intelligence/harness?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-harness-banner\"\u003e\n\u003cimg src=\"assets/harness-banner.png\" alt=\"Now includes Harness Orchestration — multi-turn coding agents with Claude Code, Codex, Gemini CLI, and OpenCode\" width=\"100%\" /\u003e\n\u003c/a\u003e\n\u003c/div\u003e\n\nAgentField is an open-source control plane that lets you build AI agents callable by any service in your stack - frontends, backends, other agents, cron jobs - just like any other API. You write agent logic in Python, Go, or TypeScript. AgentField turns it into production infrastructure: routing, coordination, memory, async execution, and cryptographic audit trails. Every function becomes a REST endpoint. Every agent gets a cryptographic identity. Every decision is traceable.\n\n\u003cdiv align=\"center\"\u003e\n\nhttps://github.com/user-attachments/assets/9fb7b1cf-26de-4b9b-9ba2-917252cc26ec\n\n\u003csub\u003e\u003cb\u003eOne prompt → a running containerized production ready multi-agent backend.\u003c/b\u003e No glue code, start using the agent API!\u003c/sub\u003e\n\n\u003c/div\u003e\n\n## Build production agents with a prompt.\n\n**Describe the system in one line. Get a production-ready multi-agent backend.** Works in Claude Code, Codex, Gemini CLI, OpenCode, Aider, Windsurf, and Cursor.\n\n```bash\ncurl -fsSL https://agentfield.ai/install.sh | bash\n```\n\nThen in your coding agent, paste any spec with /agentfield :\n\n```text\n/agentfield Build a claims processor with risk scoring, pattern detection,\nand human approval for low-confidence decisions.\n```\n\nYou get a Docker Compose stack wired up end-to-end — the agent, the control plane, and a production ready REST API endpoint you can paste and `curl` into a terminal to try it. [See it in action →](https://agentfield.ai/docs/learn/build-with-claude-code?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-prompt-to-production)\n\n## The DX you get\n\n*Best in class Python (or Go / TypeScript) DX. With least intrusive abstraction. No DSL, no YAML, no graph wiring.*\n\n```python\nfrom agentfield import Agent, AIConfig\nfrom pydantic import BaseModel\n\napp = Agent(\n    node_id=\"claims-processor\",\n    version=\"2.1.0\",# Canary deploys, A/B testing, blue-green rollouts\n    ai_config=AIConfig(model=\"anthropic/claude-sonnet-4-20250514\"),\n)\n\nclass Decision(BaseModel):\n    action: str# \"approve\", \"deny\", \"escalate\"\n    confidence: float\n    reasoning: str\n\n@app.reasoner(tags=[\"insurance\", \"critical\"])\nasync def evaluate_claim(claim: dict) -\u003e dict:\n\n    # Structured AI judgment - returns typed Pydantic output\n    decision = await app.ai(\n        system=\"Insurance claims adjuster. Evaluate and decide.\",\n        user=f\"Claim #{claim['id']}: {claim['description']}\",\n        schema=Decision,\n    )\n\n    if decision.confidence \u003c 0.85:\n        # Human approval - suspends execution, notifies via webhook, resumes when approved\n        await app.pause(\n            approval_request_id=f\"claim-{claim['id']}\",\n            approval_request_url=f\"https://internal.acme.com/approvals/claim-{claim['id']}\",\n            expires_in_hours=48,\n        )\n\n    # Route to the next agent - traced through the control plane\n    await app.call(\"notifier.send_decision\", input={\n        \"claim_id\": claim[\"id\"],\n        \"decision\": decision.model_dump(),\n    })\n\n    return decision.model_dump()\n\napp.run()\n# This single line exposes: POST /api/v1/execute/claims-processor.evaluate_claim\n# The agent auto-registers with the control plane, gets a cryptographic identity, and every\n# execution produces a verifiable, tamper-proof audit trail.\n```\n\n\u003e **What you just saw:** `app.ai()` calls an LLM and returns structured output. `app.pause()` suspends for [human approval](https://agentfield.ai/docs/build/execution/human-in-the-loop?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-human-in-the-loop). `app.call()` routes to other agents through the control plane. `app.run()` auto-exposes everything as REST. [Read the full docs →](https://agentfield.ai/docs/learn?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-read-full-docs)\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003ePrefer to scaffold by hand? (Python / Go / TypeScript / Docker)\u003c/b\u003e\u003c/summary\u003e\n\n```bash\naf init my-agent --defaults                            # Scaffold agent\ncd my-agent \u0026\u0026 pip install -r requirements.txt\naf server          # Terminal 1 → Dashboard at http://localhost:8080\npython main.py     # Terminal 2 → Agent auto-registers\n```\n\n```bash\n# Call your agent\ncurl -X POST http://localhost:8080/api/v1/execute/my-agent.demo_echo \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\"input\": {\"message\": \"Hello!\"}}'\n```\n\n```bash\n# Go\naf init my-agent --defaults --language go \u0026\u0026 cd my-agent \u0026\u0026 go run .\n\n# TypeScript\naf init my-agent --defaults --language typescript \u0026\u0026 cd my-agent \u0026\u0026 npm install \u0026\u0026 npm run dev\n\n# Docker (control plane only)\ndocker run -p 8080:8080 agentfield/control-plane:latest\n```\n\n[Deployment guide →](https://agentfield.ai/docs/reference/deploy?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-deploy) for Docker Compose, Kubernetes, and production setups.\n\n\u003c/details\u003e\n\n## How AgentField fits in your stack\n\nMost agent tools help you **write** agent logic. AgentField is what **runs** it in production — the operating layer that makes agents callable by software, durable across failures, governed by policy, and provable by audit.\n\n|  | **Frameworks**\u003cbr\u003e\u003csub\u003eLangChain · CrewAI · PydanticAI · OpenAI Agents SDK\u003c/sub\u003e | **Workflow engines**\u003cbr\u003e\u003csub\u003eTemporal · Airflow\u003c/sub\u003e | **Visual builders**\u003cbr\u003e\u003csub\u003en8n · Zapier\u003c/sub\u003e | **AgentField** |\n|---|:-:|:-:|:-:|:-:|\n| Build agent logic (prompts, tools, structured output) | ● | — | — | ● |\n| Callable production ready REST APIs out-of-box | — | ◐ | ● | ● |\n| Async + retries + webhooks | — | ● | ◐ | ● |\n| Memory scopes (global · agent · session · run) | ◐ | — | — | ● |\n| Service discovery + cross-agent calls | — | — | — | **●** |\n| Distributed agents | — | — | — | **●** |\n| Tamper-proof, verifiable audit per execution | — | — | — | **●** |\n| Harness orchestration (Claude Code · Codex · CLI) | — | — | — | **●** |\n| Identity and Access Management (IAM) for agents | — | — | — | ● |\n| Fleet observability (DAGs · metrics · traces) | — | ◐ | — | ● |\n| Multi-language SDKs (Python · Go · TypeScript) | ◐ | ● | — | ● |\n\n\u003csub\u003e● full · ◐ partial · — not the focus\u003c/sub\u003e\n\n**Use a framework when you're proving behavior.** Use AgentField when agents need to be production systems — callable by software, coordinating across services, surviving failures, and governed under audit.\n\n[Full comparison \u0026 decision guide →](https://agentfield.ai/docs/learn/vs-frameworks?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-vs-frameworks)\n\n## What You Get\n\n**Build** - Python, Go, or TypeScript. Every function becomes a REST endpoint.\n\n- **[Reasoners \u0026 Skills](https://agentfield.ai/docs/build/building-blocks/reasoners?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-reasoners)** - `@app.reasoner()` for AI judgment, `@app.skill()` for deterministic code\n- **[Structured AI](https://agentfield.ai/docs/reference/sdks/python?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-structured-ai)** - `app.ai(schema=MyModel)` → typed Pydantic/Zod output from any LLM\n- **[Harness](https://agentfield.ai/docs/build/intelligence/harness?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-harness)** - `app.harness(\"Fix the bug\")` dispatches multi-turn tasks to Claude Code, Codex, Gemini CLI, or OpenCode\n- **[Cross-Agent Calls](https://agentfield.ai/docs/build/coordination/cross-agent-calls?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-cross-agent-calls)** - `app.call(\"other-agent.func\")` routes through the control plane with full tracing\n- **[Discovery](https://agentfield.ai/docs/reference/sdks/python?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-discovery)** - `app.discover(tags=[\"ml*\"])` finds agents and capabilities across the mesh. `tools=\"discover\"` lets LLMs auto-invoke them.\n- **[Memory](https://agentfield.ai/docs/build/coordination/shared-memory?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-memory)** - `app.memory.set()` / `.get()` / `.search()` - KV + vector search, four scopes, no Redis needed\n\n**Run** - Production infrastructure for non-deterministic AI.\n\n- **[Async Execution](https://agentfield.ai/docs/build/execution/async?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-async-execution)** - Fire-and-forget with webhooks, SSE streaming, retries. No timeout limits - agents run for hours or days.\n- **[Human-in-the-Loop](https://agentfield.ai/docs/build/execution/human-in-the-loop?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-human-in-the-loop)** - `app.pause()` suspends execution for human approval. Crash-safe, durable, audited.\n- **[Canary Deployments](https://agentfield.ai/docs/learn/features?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-canary-deployments)** - Traffic weight routing, A/B testing, blue-green deploys. Roll out agent versions at 5% → 50% → 100%.\n- **[Observability](https://agentfield.ai/docs/learn/features?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-observability)** - Automatic workflow DAGs, Prometheus `/metrics`, structured logs, execution timeline.\n\n**Govern** - IAM for AI agents. Identity, access control, and audit trails - built in.\n\n- **[Cryptographic Identity](https://agentfield.ai/docs/build/governance/identity?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-crypto-identity)** - Every agent gets a W3C DID (decentralized identifier) - not a shared API key. Agents authenticate to each other the way services authenticate with mTLS, but with cryptographic signatures that travel with the agent.\n- **[Verifiable Credentials](https://agentfield.ai/docs/build/governance/credentials?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-verifiable-credentials)** - Tamper-proof receipt for every execution. Offline-verifiable: `af vc verify audit.json`.\n- **[Policy Enforcement](https://agentfield.ai/docs/build/governance/policy?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-policy-enforcement)** - Tag-based policy gates with cryptographic verification. \"Only agents tagged 'finance' can call this\" - enforced by infrastructure, not prompts.\n\n[See the full production-ready feature set →](https://agentfield.ai/docs/learn/features?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-full-features)\n\n\u003cdiv align=\"center\"\u003e\n\u003cimg src=\"assets/features-strip.png\" alt=\"90+ Production Features\" width=\"100%\" /\u003e\n\u003c/div\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003ch4 align=\"center\"\u003e▼ Click to expand full capabilities\u003c/h4\u003e\u003c/summary\u003e\n\n#### AI \u0026 LLM\n\n| Feature | How |\n|---|---|\n| Structured output (Pydantic/Zod) | `app.ai(schema=MyModel)` |\n| Multi-turn coding agents | `app.harness(\"task\", provider=\"claude-code\")` |\n| LLM auto-discovers agents and tools | `app.ai(tools=\"discover\")` |\n| Multimodal (text, image, audio) | `app.ai(\"Describe\", image_url=\"...\")` |\n| Streaming responses | `app.ai(\"...\", stream=True)` |\n| 100+ LLMs via LiteLLM | `AIConfig(model=\"anthropic/claude-sonnet-4-20250514\")` |\n| Temperature, max tokens, format | `app.ai(..., temperature=0.2)` |\n\n#### Agent Mesh \u0026 Discovery\n\n| Feature | How |\n|---|---|\n| Cross-agent calls with tracing | `app.call(\"agent.func\", input={...})` |\n| Discover agents by tag (wildcards) | `app.discover(tags=[\"ml*\"])` |\n| Discover by health status | `app.discover(health_status=\"active\")` |\n| Agent routers (namespacing) | `AgentRouter(prefix=\"billing\")` |\n| Auto context propagation | Workflow, session, actor IDs forwarded |\n| Parallel agent execution | `asyncio.gather(app.call(...), ...)` |\n| Auto-registration on startup | Service mesh with zero config |\n\n#### Execution Engine\n\n| Feature | How |\n|---|---|\n| Sync execution (REST) | `POST /api/v1/execute/{agent}.{func}` |\n| Async (fire-and-forget) | `POST /api/v1/execute/async/{agent}.{func}` |\n| Webhooks + HMAC-SHA256 signing | `AsyncConfig(webhook_url=\"...\", secret=\"...\")` |\n| SSE streaming (real-time) | `/api/v1/execute/stream/{id}` |\n| No timeout limits (hours/days) | Control plane allows unlimited duration |\n| Execution polling | `GET /api/v1/executions/{id}` |\n| Batch status checks | `POST /api/v1/executions/batch-status` |\n| Progress updates mid-execution | Intermediate payloads during long tasks |\n| Auto retries + exponential backoff | Transparent - control plane handles |\n| Backpressure + queue depth limits | Fair scheduling, circuit breakers |\n| Durable queue (PostgreSQL) | Atomic lease-based processing |\n\n#### Memory (Distributed State)\n\n| Feature | How |\n|---|---|\n| Key-value storage | `app.memory.set(key, value)` / `.get(key)` |\n| Vector search (semantic) | `app.memory.search(embedding, top_k=5)` |\n| Four scopes | Global, agent, session, run |\n| Reactive memory events | `@app.memory.on_change(\"order_*\")` |\n| Metadata filtering | Filter stored values by metadata |\n| Zero dependencies | Built into control plane - no Redis |\n\n#### Human-in-the-Loop\n\n| Feature | How |\n|---|---|\n| Durable pause/resume | `await app.pause(reason=\"...\")` |\n| Approval workflows with UI | `approval_request_url` for reviewers |\n| Configurable timeouts | `expires_in_hours=24` + auto-escalation |\n| Crash-safe state | Survives agent restarts |\n\n#### Canary Deployments \u0026 Versioning\n\n| Feature | How |\n|---|---|\n| Traffic weight routing | 5% → 50% → 100% rollouts |\n| A/B testing | 50/50 splits with `X-Routed-Version` |\n| Blue-green deployments | Instant weight switch, zero downtime |\n| Per-version health tracking | Unhealthy versions auto-removed |\n| Agent lifecycle states | pending → starting → ready → degraded → offline |\n\n#### Identity \u0026 Governance\n\n| Feature | How |\n|---|---|\n| Cryptographic identity per agent | Auto-generated W3C DID + Ed25519 keys |\n| Verifiable Credentials | Tamper-proof receipt per execution |\n| Offline VC verification | `af vc verify audit.json` |\n| Tag-based access policies | ALLOW/DENY rules on caller → target tags |\n| Cryptographically signed requests | Ed25519 signatures on cross-agent calls |\n| VC hierarchy (3 tiers) | Platform → Node → Function control |\n| Agent notes (audit log) | `app.note(\"Decision\", tags=[\"critical\"])` |\n| Non-repudiation | Cryptographic proof of actions |\n| Permission request workflows | Auto-created when access denied |\n\n#### Observability \u0026 Fleet Management\n\n| Feature | How |\n|---|---|\n| Automatic DAG visualization | Workflow graphs in dashboard |\n| Prometheus metrics | `/metrics` out of the box |\n| Structured JSON logging | Automatic from SDK |\n| Execution timeline | Chronological decision trace |\n| Health checks (K8s-ready) | `/health`, `/ready` endpoints |\n| Correlation IDs | `X-Workflow-ID`, `X-Execution-ID` |\n| Workflow DAG API | `GET /api/v1/workflows/{id}/dag` |\n| Agent heartbeat monitoring | Auto health status transitions |\n\n#### Harness (Multi-turn Coding Agents)\n\n| Feature | How |\n|---|---|\n| 4 providers | Claude Code, Codex, Gemini CLI, OpenCode |\n| Schema-constrained output | `schema=ResultModel` (Pydantic/Zod) |\n| Cost capping | `max_budget_usd=3.0` |\n| Turn limiting | `max_turns=100` |\n| Tool access control | `tools=[\"Read\", \"Write\", \"Bash\"]` |\n| Environment injection | `env={\"KEY\": \"value\"}` |\n| System prompt override | `system_prompt=\"...\"` |\n| Multi-layer output recovery | Cosmetic repair → retry → full retry |\n\n#### Connector API (Fleet Management)\n\n| Feature | How |\n|---|---|\n| Remote agent management | `/connector/reasoners` |\n| Version traffic control | `/connector/.../weight` |\n| Bearer token auth | `AGENTFIELD_CONNECTOR_TOKEN` |\n| Air-gapped deployment | Outbound WebSocket only |\n\n#### Developer Experience\n\n| Feature | How |\n|---|---|\n| CLI scaffolding | `af init my-agent --defaults --language python\\|go\\|typescript` |\n| Local dev with dashboard | `af server` → http://localhost:8080 |\n| Hot reload | `af dev` auto-detects changes |\n| Auto-REST from decorators | Every `@app.reasoner()` → `POST /api/v1/execute/...` |\n| Python, Go, TypeScript SDKs | Native patterns per language |\n| MCP server integration | `af add --mcp --url \u003cserver\u003e` |\n| Config storage API | `POST /api/v1/configs/:key` - database-backed |\n| Docker + Kubernetes ready | Stateless control plane, horizontal scaling |\n\n[Explore all features in detail →](https://agentfield.ai/docs/learn/features?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-explore-features)\n\n\u003c/details\u003e\n\n## Built With AgentField\n\n\u003ctable\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/github/swe-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-swe-af-repo\"\u003e\n        \u003cimg src=\"assets/examples/autonomous-engineering-team.png\" alt=\"Autonomous Engineering Team\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eAutonomous Engineering Team\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eOne API call spins up PM, architect, coders, QA, reviewers - hundreds of coordinated agents that plan, build, test, and ship.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/github/swe-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-swe-af-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/github/deepresearch/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-deepresearch-repo\"\u003e\n        \u003cimg src=\"assets/examples/deep-research-engine.png\" alt=\"Deep Research Engine\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eDeep Research Engine\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eRecursive research backend. Spawns parallel agents, evaluates quality, generates deeper agents, and recurses -10,000+ agents per query.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/github/deepresearch/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-deepresearch-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/github/mongodb/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-mongodb-repo\"\u003e\n        \u003cimg src=\"assets/examples/reactive-database-layer.png\" alt=\"Reactive MongoDB Intelligence\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eReactive MongoDB Intelligence\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eAtlas Triggers + agent reasoning. Documents arrive raw and leave enriched - risk scores, pattern detection, evidence chains.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/github/mongodb/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-mongodb-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/github/sec-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-sec-af-repo\"\u003e\n        \u003cimg src=\"assets/examples/ai-security-auditor.png\" alt=\"Autonomous Security Audit\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eAutonomous Security Audit\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003e250 coordinated agents trace every vulnerability source-to-sink and adversarially verify each finding. Confirmed exploits, not pattern flags.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/github/sec-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-sec-af-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/github/cloudsecurity/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_content=cloudsec\u0026utm_id=github-readme-cloudsec-repo\"\u003e\n        \u003cimg src=\"assets/examples/cloud-security.png\" alt=\"CloudSecurity AF\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eCloudSecurity AF\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eAI-native cloud infrastructure security scanner that performs shift-left attack path analysis directly from IaC, prioritizing the most dangerous risk chains before deployment.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/github/cloudsecurity/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_content=cloudsec\u0026utm_id=github-readme-cloudsec-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/pr-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-pr-af-repo\"\u003e\n        \u003cimg src=\"assets/examples/agentic-pr-reviewer.png\" alt=\"Agentic PR Reviewer\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eAgentic PR Reviewer\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eBuilds a custom review strategy for every PR - spawns parallel reviewer agents with runtime-crafted prompts, adversarially challenges its own findings, and posts evidence-grounded inline comments.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/pr-af/?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-pr-af-repo\"\u003eView project →\u003c/a\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n\u003c/table\u003e\n\n[See all examples →](https://www.agentfield.ai/examples?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-see-all-examples)\n\nBuilt something with AgentField? [Submit your project to be featured on the examples page](https://github.com/Agent-Field/agentfield/issues/new?template=community-project.md).\n\n## See It In Action\n\n\u003cdiv align=\"center\"\u003e\n\u003cimg src=\"assets/UI.png\" alt=\"AgentField Dashboard\" width=\"100%\" /\u003e\n\u003cbr/\u003e\n\u003csub\u003eReal-time workflow DAGs · Execution traces · Agent fleet management · Audit trails\u003c/sub\u003e\n\u003c/div\u003e\n\n## Architecture\n\n\u003cdiv align=\"center\"\u003e\n\u003cimg src=\"assets/arch.png\" alt=\"AgentField Architecture\" width=\"100%\" /\u003e\n\u003c/div\u003e\n\nThe control plane is a stateless Go service. Agents connect from anywhere - your laptop, Docker, Kubernetes. They register capabilities, the control plane routes calls between them, tracks execution as DAGs, and enforces policies. [Full architecture docs →](https://agentfield.ai/docs/learn/architecture?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-architecture)\n\n## Learn More\n\nThe thinking behind AgentField - essays on AI backends, harness orchestration, and the infrastructure production agents actually need.\n\n\u003ctable\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/blog/what-is-harness-orchestration?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-what-is-harness-orchestration\"\u003e\n        \u003cimg src=\"assets/blog/what-is-harness-orchestration.webp\" alt=\"What is harness orchestration?\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eWhat is harness orchestration?\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eThe atomic unit of intelligence is climbing from the model call to the autonomous harness - and what changes when it does.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/blog/what-is-harness-orchestration?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-what-is-harness-orchestration\"\u003eRead post →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/blog/harness-as-black-box?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-harness-black-box\"\u003e\n        \u003cimg src=\"assets/blog/harness-black-box.webp\" alt=\"Part 1: The Black Box\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003ePart 1: The Black Box\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eTreating harnesses like Claude Code and Codex as autonomous, embodied, persistent computational entities.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/blog/harness-as-black-box?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-harness-black-box\"\u003eRead post →\u003c/a\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/blog/harness-as-membrane?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-harness-membrane\"\u003e\n        \u003cimg src=\"assets/blog/harness-membrane.webp\" alt=\"Part 2: Engineering the Membrane\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003ePart 2: Engineering the Membrane\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eShaping the boundary surface of a harness across four engineerable dimensions: workspace, drift, verifier placement, and recovery budget.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/blog/harness-as-membrane?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-harness-membrane\"\u003eRead post →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/blog/ai-backend?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-ai-backend\"\u003e\n        \u003cimg src=\"assets/blog/ai-backend.webp\" alt=\"The AI Backend\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eThe AI Backend\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eOur thesis: in five years every serious software company will run an AI backend - a reasoning layer that makes the decisions that used to be hardcoded.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/blog/ai-backend?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-ai-backend\"\u003eRead post →\u003c/a\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n      \u003ca href=\"https://agentfield.ai/blog/iam-ai-backends?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-iam\"\u003e\n        \u003cimg src=\"assets/blog/iam-ai-backends.png\" alt=\"IAM for AI Backends\" /\u003e\n      \u003c/a\u003e\n      \u003cbr/\u003e\n      \u003cb\u003eIAM for AI Backends\u003c/b\u003e\n      \u003cbr/\u003e\n      \u003csub\u003eAgents need identity, not API keys - how decentralized identifiers and verifiable credentials make agent-to-agent delegation auditable and accountable.\u003c/sub\u003e\n      \u003cbr/\u003e\u003cbr/\u003e\n      \u003ca href=\"https://agentfield.ai/blog/iam-ai-backends?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-blog-iam\"\u003eRead post →\u003c/a\u003e\n    \u003c/td\u003e\n    \u003ctd align=\"center\" width=\"50%\"\u003e\n    \u003c/td\u003e\n  \u003c/tr\u003e\n\u003c/table\u003e\n\n### Documentation\n\n- **[vs Agent Frameworks](https://agentfield.ai/docs/learn/vs-frameworks?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-vs-frameworks)** - How AgentField compares to LangChain, CrewAI, and workflow engines\n- **[Full Documentation](https://agentfield.ai/docs/learn?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-full-docs)**\n\n## Community\n\n\u003cdiv align=\"center\"\u003e\n\n[![Discord](https://img.shields.io/badge/Join%20Discord-d4a24a?style=for-the-badge\u0026logo=discord\u0026logoColor=white)](https://discord.gg/aBHaXMkpqh)\n[![Twitter](https://img.shields.io/badge/Follow%20on%20X-0c0b09?style=for-the-badge\u0026logo=x\u0026logoColor=white)](https://x.com/agentfield_ai)\n\n**[GitHub Issues](https://github.com/Agent-Field/agentfield/issues)** · **[Documentation](https://agentfield.ai/docs/learn?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-community-docs)** · **[Examples](https://agentfield.ai/docs/learn/examples?utm_source=github-readme\u0026utm_campaign=github-readme\u0026utm_id=github-readme-community-examples)**\n\n\u003c/div\u003e\n\n## License\n\n[Apache 2.0](LICENSE)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fagent-field%2Fagentfield","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fagent-field%2Fagentfield","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fagent-field%2Fagentfield/lists"}