當 agent 代表人或服務完成一項工作，真正需要說明的是三件事：誰在行動、代表誰、這次行動被允許做什麼。把它們分開，授權和後續核驗才有明確對象。

文章

Agent · Articles

Agent — ArcBlock builds decentralized identity, user-owned data, Blocklet infrastructure, and AI-native application systems.

OpenAI's Agents API is a harness. ARC is a new kind of computer for agents

OpinionSep 11, 2026

OpenAI's Agents API is a harness. ARC is a new kind of computer for agents

OpenAI's Agents API turns the Codex loop into a hosted API: sessions, compaction, tools, sandboxes, plugins. That layer is real. It is still not a computer with identity, durable state, and an owner who can walk away with it. ARC attaches at the seam they already documented.

OpinionAug 21, 2026

Equalized capability does not automatically concentrate

When AI equalizes the ability to build, control does not automatically tighten. Many small units and home tools can coexist; super-individuals and giant canteens are not the default.

OpinionAug 21, 2026

If you still picture the old software and the old software company, you will misread this shift

Many people still imagine software as slow, expensive craft, and software companies as hiring engines. Both have changed. Building it is no longer the main bottleneck.

ProductAug 20, 2026

Where your AI spend goes: an introduction to Agent Observability

Every model call an agent makes is written to a ledger. This is where those ledgers live, which of the three pages answers which question, whose data you are allowed to see, and the two boundaries you have to know before you read a number off any of them.

OpinionAug 18, 2026

After Software: Collaboration, Production, and How Value Gets Shared

Making software no longer requires making a company first. When productive capacity is scattered across people and AI, collaboration, ownership, and value need a new organization.

NewsAug 14, 2026

ARC Now Supports the MCP 2026-07-28 Specification

ARC already spoke MCP. This upgrade aligns that surface with MCP 2026-07-28: stateless streamable HTTP and standard OAuth discovery so third-party agents can reach Blocklets without a special bridge.

OpinionAug 10, 2026

The Infrastructure Came First. The Residents Are Arriving.

Not a pivot from blockchain to AI. The harder claim is simpler: roads and utilities can be built early, but a digital city only comes alive when residents who can work around the clock walk in.

OpinionAug 10, 2026

Builders Change. Your Property Stays.

Most people will not manage API keys. They will hire agents. The real question is not which model is smarter, but whose ground the worker stands on when the job is done.

EngineeringAug 10, 2026

What `arc skill` Is Today (and What It Is Not)

arc skill ships a small catalog of built-in how-to booklets and can drop them into your agent skill folders. That is useful—and incomplete next to gh skill or AFS’s own .skills trees. Here is the plain map, a short how-to, and the open design gap.

OpinionAug 9, 2026

Agents don't need a better sandbox. They need a computer of their own.

Many companies are building compute for agents: sandboxes, VMs, even digital workers. They answer where untrusted code should run. What agents still lack is a machine that can last: identity, state, constrained authority, verifiable history, and the right to leave. We call that machine ARC.

ExplainerAug 9, 2026

When a writing agent should leave the page without a figure

Writing agents on this site plan figures at outline time, then run kind-specific skills. Zero figures is valid. When a figure earns its place, the pipeline leaves a structured source, a rendition, and a text fallback.

OpinionAug 8, 2026

Don't Take Your Project to an Agent Factory. Invite Agents Into Your Own Realm.

The important question is not whether an AI model runs locally. It is whether a project remains a place where its people can see, govern, and recover its context, authority, evidence, and change.

OpinionAug 7, 2026

A Useful Token Must First Stand for Something of Real Value

AI services already use credits to meter storage, compute, and API calls. As Agents begin to act across systems, the important work is not giving those units a fashionable name. It is making the rights, authority, and usage they represent possible to verify and reconcile.

EngineeringJul 31, 2026

Small World: Every Agent Lives in Its Own Projection

AFS has a design principle we call Small World: every agent, every app, every process sees not the global world, but a small world projected specifically for it. Anything outside that world isn't "access denied". It simply doesn't exist. Here's where the idea comes from, why it's more than a security mechanism, and how to try it yourself.

NewsJul 9, 2025

Agent · Articles

OpenAI's Agents API is a harness. ARC is a new kind of computer for agents

Equalized capability does not automatically concentrate

If you still picture the old software and the old software company, you will misread this shift

Where your AI spend goes: an introduction to Agent Observability

After Software: Collaboration, Production, and How Value Gets Shared

ARC Now Supports the MCP 2026-07-28 Specification

The Infrastructure Came First. The Residents Are Arriving.

Builders Change. Your Property Stays.

What `arc skill` Is Today (and What It Is Not)

Agents don't need a better sandbox. They need a computer of their own.

When a writing agent should leave the page without a figure

Don't Take Your Project to an Agent Factory. Invite Agents Into Your Own Realm.

A Useful Token Must First Stand for Something of Real Value

Small World: Every Agent Lives in Its Own Projection

ArcBlock 前往 SeattleJS：展示多代理系統的未來

AI 代理需要錢包——為什麼去中心化身份比以往任何時候都更重要

ArcBlock 加入去中心化 AI 代理聯盟 (DAIAA)

如何建立一個 AI 驅動的會議摘要與行動項代理

使用 AIGNE 打造專屬的學習計畫代理 (Agent)

AIGNE v0.4.20:支援 Agent 快取、排程任務與串流回應

AIGNE 0.2 正式發布:支援多重代理與決策型代理