본문으로 건너뛰기
cccskills
무료GitHub에 공개

agents-sdk

Build stateful Agents on Cloudflare Workers with current, source-verified SDK APIs. Covers durable state, RPC, chat, schedules, queues, fibers, Workflows, MCP, and channels while keeping authorization and external effects application-owned. NOT for assuming SDK helpers grant application authorization or effect approval.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

연동 도구Cloudflare
설치 방법 보기

포함된 파일(21)

  • SKILL.md4.6 KB
  • references/browse-the-web.md5.7 KB
  • references/callable.md3.3 KB
  • references/client-sdk.md8.9 KB
  • references/codemode.md3.3 KB
  • references/configuration.md4.4 KB
  • references/durable-execution.md4.4 KB
  • references/email.md10.4 KB
  • references/evidence-scope.md1.3 KB
  • references/human-in-the-loop.md3.1 KB
  • references/mcp.md12.9 KB
  • references/observability.md2.2 KB
  • references/queue-retries.md3.8 KB
  • references/routing.md7.6 KB
  • references/server-driven-messages.md10.7 KB
  • references/state-scheduling.md13.8 KB
  • references/streaming-chat.md15.8 KB
  • references/think.md5.8 KB
  • references/voice.md1.7 KB
  • references/webhooks-push.md4.7 KB
  • references/workflows.md5.1 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Cloudflare Agents SDK

Prefer the installed package and current Cloudflare Agents documentation over remembered signatures. This bundle was reconciled on 2026-09-24; examples are illustrative and were not package-typechecked or deployed.

Runtime and effect boundary

Cloudflare Agents provides Durable Object-backed identity, local SQLite state, connections, scheduling, and durable execution primitives. The application must configure principal authentication and resource authorization, make provider retries safe, and verify external outcomes separately from transcript/event persistence. Durable Object instance identity does not establish user authority.

flowchart LR
  I[Authorized input] --> A[Agent runtime]
  A --> S[(Durable state or work record)]
  A --> P[Scoped application operation]
  P --> X[External provider]
  X --> R[Readback or effect receipt]
  R --> S

Pick a primitive

NeedStart hereRequired application boundary
Stateful real-time instancestate, routing, client SDKauthenticate and authorize instance access
Browser commandcallable methodsvalidate input, idempotency, outcome status
Scheduled/queued callbackstate scheduling, queuesduplicate, retry, and late-delivery behavior
Recovery from DO evictiondurable executioncomplete checkpoint and recovery/compensation policy
Long-running multi-step work or approvalWorkflowsbind request/action/approval and repeated progress
Chat/server turnstreaming chat, server-driven messagestranscript persistence differs from effect completion
External webhook/pushwebhooks/pushverify source, dedupe, current consent, effect receipt
sequenceDiagram
  participant C as caller
  participant A as Agent
  participant D as durable record
  participant X as external system
  C->>A: authorized request and operation key
  A->>D: record accepted intent
  A->>X: attempted operation
  Note over A,X: timeout, disconnect, or eviction can leave outcome unknown
  A->>X: query by original operation identity when supported
  X-->>A: definitive outcome or unknown
  A->>D: inspect receipt or retain unknown
  alt independently confirmed
    A->>D: record completed effect
  else not confirmed
    A->>D: pending reconciliation under original identity
  end

Navigation

Core runtime

Durable/background execution

  • Queues and retries: FIFO queue, retry classification, unknown result handling.
  • Fibers: checkpoints, recovery, startFiber, cooperative cancellation.
  • Workflows: durable steps, external waits, approval and progress.
  • Observability: diagnostics, Tail Workers, correlation.

Channels and harnesses

Source and validation limit

The diagrams show a design boundary, not a runtime trace. Each reference names the official sources and the date it was checked; evidence scope describes the validation boundary. No installation, typecheck, Worker, Durable Object, Cloudflare service, provider, approval, notification, or browser action was run.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Expert in 2000s-era music visualization (Milkdrop, AVS, Geiss) and modern WebGL implementations. Specializes in Butterchurn integration, Web Audio API AnalyserNode FFT data, GLSL shaders for audio-reactive visuals, and psychedelic generative art. Activate on "Milkdrop", "music visualization", "WebGL visualizer", "Butterchurn", "audio reactive", "FFT visualization", "spectrum analyzer". NOT for simple bar charts/waveforms (use basic canvas), video editing, or non-audio visuals.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

Expert legal research agent for finding and scraping expungement data state by state. Knows authoritative sources, URL patterns, Firecrawl configuration, and 2026 legal landscape.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

Apply crisis decision-making research to agent routing, uncertainty triage, and coordination failure analysis in time-pressured systems. Use when diagnosing handoff failures, analytical paralysis, or expert judgment under incomplete information. NOT for routine coding, simple CRUD design, or static single-agent tasks with complete information.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

Use for insight, reframing, contradiction, impasse, and anomaly-driven problem solving when execution effort no longer helps. NOT for routine optimization, error correction, or well-specified tasks with known solution paths.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

Apply cognitive task analysis to expert work that depends on perceptual cues, branching judgment, and recurring monitoring loops. Use when decomposing expert capability into agent structure, simulation design, or validation interviews. NOT for ordinary step-by-step SOP capture or simple pipelines with no tacit cue layer.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

curiositech/port-daddy32026년 10월 18일 업데이트

curiositech의 스킬 모두 보기

이 스킬의 문제 신고