简报·YT
← 返回 GitHub 动态
Significant-Gravitas/AutoGPT · 📈 热门项目 已完成

Significant-Gravitas/AutoGPT autogpt-platform-beta-v0.8.2

2026-09-30 05:25 · ⭐ 18.8万 · GitHub Release 原文 ↗
#智能体平台 #权限审批 #版本发布

Significant-Gravitas/AutoGPT autogpt-platform-beta-v0.8.2

项目简介

AutoGPT 是较早探索"自主 AI 智能体"方向的开源项目，如今已演进为一个可在浏览器中使用、面向普通用户的 AI 智能体平台（AutoGPT Platform）：用户可以用自然语言编排工作流、雇佣"Expert"（专家智能体）、连接外部集成与工具，让智能体自动完成搜索、读写文件、调用 API 等任务。它值得关注，是因为它在"智能体自主执行"与"用户可控审批"之间做了大量工程化尝试，并把权限门控、成本控制、凭据隔离这些生产级问题当作核心设计目标。

一句话概括

本版本围绕 AutoPilot 的行动门控（action gating）与审批体系做了系统性重构——引入 Ask First / Auto / Unsupervised 三种模式、按对话/专家/团队分级的审批规则、由确定性组件（Jev）而非 LLM 做关键判断，同时补齐了新的搜索/集成能力、扩充 LLM 模型目录并修复了大量定价、鉴权与前端问题。

更新要点

1. AutoPilot 行动门控与审批体系（本版本主线）

2. 判断逻辑从 LLM 转向确定性组件 + 性能数据

3. 新增能力与集成

4. LLM 模型目录更新

5. UI/UX 改进

6. 主要 Bug 修复

7. 依赖与 CI/CD

8. Breaking Change

Release Notes 中未声明任何 breaking change。升级方式：

git pull origin autogpt-platform-beta-v0.8.2

9. 贡献者

新贡献者：@Xxx91n、@JohnnyWilson16；常驻贡献者包括 @Pwuts、@ntindle、@0ubbe、@kcze、@Abhi1992002、@Bentlybro、@Torantulino、@itsababseh、@dependabot[bot]。

▸ Release Notes 英文原文（11133 字符）
## 🔥 What's New? ### New Features - **#14209** - AutoPilot action gating modes - Ask First, Auto, Unsupervised (by @Pwuts) - **#14817** - Workflows started from AutoPilot pause before an irreversible action; the irreversible marker renamed and widened (by @Pwuts) - **#14832** - AutoPilot carries on while a held call awaits review: a queue of cards per chat, late results on approval (by @Pwuts) - **#14837** - AutoPilot runs read-only and workspace blocks and workflows without asking, and names the block or workflow on every card (by @Pwuts) - **#14839** - AutoPilot decides MCP calls on each server's effect map and asks on first use of any other tool (by @Pwuts) - **#14834** - Outside reads are judged before the model sees them; a read carrying instructions is held for the user (by @Pwuts) - **#14841** - Paid reads ask once the chat is over its spend ceiling (by @Pwuts) - **#14645** - Credential swap proxy that every E2B box egresses through (by @kcze) - **#14792** - Add AnySearch search, parallel-search and extract blocks (by @Xxx91n) - **#14775** - Support connected integrations in copilot @ mentions (by @Abhi1992002) - **#14695** - Add browser notification settings to the account page (by @kcze) - **#14701** - Tell the main AutoPilot chat about its computer (by @kcze) - **#14849** - Hiring an expert returns at once; its skills and workflows install in the background (by @Pwuts) - **#14853** - AutoPilot's approval requests belong to their chat, not a made-up graph execution (by @Pwuts) - **#14878** - Credit history and the cost log name the AutoPilot chat, not a made-up graph execution (by @Pwuts) - **#14931** - Approval rules can hold for one chat, one Expert, or the whole team (by @Pwuts) - **#14980** - Approval cards for bare tools offer "Approve from now on" and "Let Otto judge from now on" (by @Pwuts) - **#14997** - Trust reads of memories, installed skills and owned workspace paths (by @Pwuts) - **#14999** - Review held AutoPilot calls in a dialog (by @Pwuts) - **#14858** - Add clay avatar picker and live AI generation (by @0ubbe) - **#14901** - Include public expert pages in a maintained sitemap (SECRT-2750) (by @Abhi1992002) - **#14945** - Publish the skills catalog on deploy and update installed copies in place (by @Torantulino) - **#14859** - Add PostHog MCP and complete integration icons (by @ntindle) - **#14971** - Add OpenSEO MCP, including self-hosted instances (by @ntindle) - **#14891** - A held call names the things its IDs point at, with links (by @Pwuts) - **#14494** - Require a short CTA-style sub_heading on marketplace listings, with a backfill script (by @Pwuts) - **#14833** - Single PostHog event list and name constants (by @kcze) - **#15040** - Keep subscription status and lifecycle dates on the PostHog person (by @kcze) - **#15026** - Backfill existing subscribers into MailerLite and add a trial group (by @Bentlybro) ### Enhancements - **#14867** - Speed up expert hires by batching skill installs and skipping repeat embeds and scans (by @ntindle) - **#14765** - Put experts first when hire-experts is on (by @0ubbe) - **#14928** - Tuned action-supervisor rubric for Haiku (98% recall, 43% needless holds on holdout) (by @Pwuts) - **#14932** - Jev decides the action supervisor's calls; the LLM only explains an ask (by @Pwuts) - **#15006** - Jev decides the content judge, Sonnet only quotes a hold (by @Pwuts) - **#14875** - Preserve main content in web_fetch and warn on JS-only shells (by @JohnnyWilson16) - **#14883** - Add xAI Grok 4.7 to the catalog (by @ntindle) - **#14903** - Add GLM-5.3 to LLM catalog (by @ntindle) - **#14943** - Add Claude Opus 4.8 (claude-opus-4-8) (by @ntindle) - **#14988** - Add Xiaomi MiMo-V2.6-Pro and MiMo-V2.6-Flash to LLM catalog (by @ntindle) - **#15014** - Add Claude Sonnet 5.5 and default the standard tier to it (by @ntindle) - **#14949** - Promote gpt-6-luna to fast/standard default over gpt-5.6-luna (by @ntindle) ### UI/UX Improvements - **#14885** - AutoPilot action gating card and queue design update (by @Pwuts) - **#14876** - Offer the free trial on the platform paywall (by @0ubbe) - **#14934** - Match artifact panel to Claude reference (by @0ubbe) - **#14933** - Reduce the expert avatar creator to one AI avatar (by @0ubbe) - **#15002** - Expert card and chat chip polish (by @0ubbe) - **#15008** - Show avatar defaults before optional generation (by @0ubbe) - **#15029** - Show chat document count and expert files (by @0ubbe) - **#15033** - Show expert documents as Files-page rows (by @0ubbe) - **#15041** - Drop section title icons and unify heading sizes (by @0ubbe) - **#14935** - Guide existing users to workflows in Otto (by @Torantulino) - **#14948** - Serve every Expert from the managed Clay & Rock library and constrain avatar generation to the design system (by @itsababseh) - **#14994** - Compress Clay and Rock avatar assets (by @0ubbe) ### Bug Fixes - **#14854** - Give the flag tests a consent answer after the master back-merge (by @Pwuts) - **#14777** - Keep a mid-turn follow-up queued until the backend drains it (by @Abhi1992002) - **#14493** - Fall back gracefully when an avatar or listing image URL is dead (by @Pwuts) - **#14855** - Repoint copilot advanced-tier Opus alias to claude-opus-5.5 (by @ntindle) - **#14856** - Correct deepseek-v4.1-flash pricing drift (again) (by @ntindle) - **#14894** - Send the platform PKCE verifier on Google token exchange (by @itsababseh) - **#14897** - Send the platform PKCE verifier on Google token exchange (prod replica of #14894) (by @Pwuts) - **#14896** - Bump claude-agent-sdk so the bundled Claude Code accepts Opus 5.5 (by @Pwuts) - **#14899** - Save thumbs up/down feedback as Langfuse scores (by @ntindle) - **#14880** - Stop the Sentry scrubber silently dropping error events (by @Bentlybro) - **#14921** - Finish skipped nodes instead of leaving them QUEUED (by @ntindle) - **#14924** - Ask for a file instead of silently skipping an empty picker (by @ntindle) - **#14902** - Make public expert profiles readable by search engines (SECRT-2749) (by @Abhi1992002) - **#14907** - Keep the Drive file you picked when you save your agent (by @ntindle) - **#14904** - Correct OpenRouter price drift for 3 models (by @ntindle) - **#14944** - Correct OpenRouter price drift for inclusionai/ling-3.0-flash-vl (by @ntindle) - **#14954** - Update pre-catalog copies and harden the marketplace copy reconcile (by @ntindle) - **#14884** - Stop update_skills deleting skills the client never saw (by @Bentlybro) - **#14973** - Settle a held call's inline step once its approval runs (by @Pwuts) - **#14975** - Say so when a chat link can't be opened on this account (by @Pwuts) - **#14974** - An approval's continuation reply streams into the chat that approved it (by @Pwuts) - **#14979** - A hire's onboarding card is answerable from the team-page chat (by @Pwuts) - **#14959** - Correct deepseek/deepseek-v4.1-flash pricing drift in catalog.py (by @ntindle) - **#14989** - Correct deepseek-chat and kimi-k2.6 pricing drift in catalog.py (by @ntindle) - **#14982** - A held call's chain row shows the answer at the click (by @Pwuts) - **#15005** - A held read names the acting Expert as its reader; Otto stays the judge (by @Pwuts) - **#15007** - Give the content judge room to think and one retry on an empty answer (by @Pwuts) - **#14996** - Show a held read's reason in plain words, as the chat does (by @Pwuts) - **#15015** - Load the block registry before taking runs (by @Bentlybro) - **#15017** - Correct Grok 4.7 OpenRouter display price drift (by @ntindle) - **#15012** - Sonnet's content judge holds agent-addressed text whatever its source (by @Pwuts) - **#15022** - Keep one-time secrets out of session replays (by @kcze) - **#15024** - Show the sign-in card instead of holding it as an outside read (by @Pwuts) - **#15030** - Send CSP frame-ancestors and Referrer-Policy on every route (by @0ubbe) - **#15038** - Restore pending expert onboarding in new chats (by @0ubbe) - **#15034** - Start expert onboarding after raise and from the team chat panel (by @0ubbe) ### Dependencies & Maintenance - **#14560** - Bump actions/stale from 10 to 11 (by @dependabot[bot]) - **#14657** - Bump boring-avatars from 1.11.2 to 2.0.4 in /autogpt_platform/frontend (by @dependabot[bot]) - **#14658** - Bump @vercel/analytics from 1.5.0 to 2.0.1 in /autogpt_platform/frontend (by @dependabot[bot]) - **#14652** - Bump shepherd.js from 14.5.1 to 15.3.0 in /autogpt_platform/frontend (by @dependabot[bot]) - **#14656** - Bump motion from 12.38.0 to 13.2.0 in /autogpt_platform/frontend (by @dependabot[bot]) - **#14735** - Bump exa-py from 1.16.1 to 2.21.0 in /autogpt_platform/backend (by @dependabot[bot]) - **#14653** - Bump framer-motion from 12.23.24 to 13.3.0 in /autogpt_platform/frontend (by @dependabot[bot]) - **#14845** - Bump ruff from 0.16.7 to 0.16.8 in /autogpt_platform/autogpt_libs in the development-dependencies group (by @dependabot[bot]) - **#14937** - Remove unused dependencies (by @0ubbe) - **#14920** - Delete the graphs and store listings that tests create (by @ntindle) - **#14923** - Stop parallel test runs clobbering each other's docker stack (by @ntindle) - **#14969** - Put the single-container image's code layers last (by @ntindle) - **#14998** - Drop a held read's bytes from the attention feed (by @Pwuts) ### CI/CD Improvements - **#14870** - Run the full-stack e2e job on a free runner (by @Bentlybro) - **#14877** - Speed up the full-stack e2e job on the free runner (by @Bentlybro) - **#14892** - Cache the e2e Docker layers in GHCR instead of the Actions cache (by @Bentlybro) - **#14947** - Let /reapprove restore the PR reviewer bot's dismissed approvals (by @ntindle) - **#14965** - Stop an out-of-quota Copilot review from blocking every merge (by @ntindle) - **#14964** - Default a manual dev deploy to the dev branch, not master (by @ntindle) - **#14966** - Stop pinning the catalog's expert skill lists in backend tests (by @ntindle) - **#15013** - Let single-container PR builds read the GHCR layer cache (by @Bentlybro) --- ## 🎉 Thanks to Our Contributors! A huge shoutout to everyone who contributed to this release. Special thanks to our new contributors: - @Xxx91n - @JohnnyWilson16 As well as our returning contributors: - @Pwuts - @ntindle - @0ubbe - @kcze - @Abhi1992002 - @Bentlybro - @Torantulino - @itsababseh - @dependabot[bot] --- ## 📥 How to Get This Update To update to this version, run: ```bash git pull origin autogpt-platform-beta-v0.8.2 ``` Or download it directly from the [Releases page](https://github.com/Significant-Gravitas/AutoGPT/releases). For a complete list of changes, see the [Full Changelog](https://github.com/Significant-Gravitas/AutoGPT/compare/autogpt-platform-beta-v0.8.1...autogpt-platform-beta-v0.8.2). --- ## 📝 **Feedback and Issues** If you encounter any issues or have suggestions, please [join our Discord](https://discord.gg/autogpt) and let us know!
▸ 主项目介绍 · Significant-Gravitas/AutoGPT
▸ README 原文（同步于 2026-07-19 13:27）
<p align="center"> <a href="https://platform.agpt.co/signup?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=hero_banner"> <img src="docs/home/.gitbook/assets/Banner_image.png" alt="AutoGPT" width="100%" /> </a> </p> <h1 align="center">AutoGPT — AI agents that finish the work</h1> <p align="center"> <strong>Get 10 hours back every week.</strong><br /> Describe what you want done. AutoGPT builds the agent, runs it, and reports back. </p> <p align="center"> <a href="https://platform.agpt.co/signup?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=header_get_started"><strong>Get started</strong></a> · <a href="https://platform.agpt.co/tour?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=tour_header">Tour</a> · <a href="https://agpt.co/pricing?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=header_pricing">Pricing</a> · <a href="https://docs.agpt.co">Docs</a> · <a href="https://discord.gg/autogpt">Discord</a> · <a href="#self-host-autogpt">Self-host</a> </p> <p align="center"> <a href="https://discord.gg/autogpt"> <img src="https://img.shields.io/badge/Discord-join-5865F2?logo=discord&logoColor=white" alt="Join Discord" /> </a> <a href="https://docs.agpt.co"> <img src="https://img.shields.io/badge/Docs-read-0A7AFF?logo=gitbook&logoColor=white" alt="Read the docs" /> </a> </p> --- ## The open-source platform for AI agents AutoGPT lets you build, deploy, and run AI agents that carry out complete workflows. Describe an outcome in plain English or shape every step in the visual builder, then run the agent on demand, on a schedule, or from a trigger. **185,000+ GitHub stars. Cited by:** | | | |---|---| | “Next frontier of prompt engineering imo: ‘AutoGPTs’.” | **Andrej Karpathy**, founding member of OpenAI | | “If you have a phone you can run AutoGPT. You don't even need to learn how to code.” | **Amjad Masad**, co-founder & CEO of Replit | | “AutoGPT might be the next big step in AI.” | **Lior Alexander**, CEO of AlphaSignal | --- ## Four surfaces, one platform <table> <tr> <td width="50%" align="center" valign="top"> <a href="https://agpt.co/product/autopilot/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=autopilot"> <img src="docs/content/imgs/readme/autogpt_autopilot_chat.jpg" alt="AutoPilot chat creating an AutoGPT agent" /> </a> <br /><strong><a href="https://agpt.co/product/autopilot/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=autopilot">AutoPilot</a></strong> <br />Describe the job in plain English and turn the conversation into a working agent. </td> <td width="50%" align="center" valign="top"> <a href="https://agpt.co/product/agents/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=agents"> <img src="docs/content/imgs/readme/autogpt_agent_dashboard.jpg" alt="Agents dashboard showing statuses, runs, and costs" /> </a> <br /><strong><a href="https://agpt.co/product/agents/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=agents">Agents</a></strong> <br />See every agent, run, cost, and action that needs your attention. </td> </tr> <tr> <td width="50%" align="center" valign="top"> <a href="https://agpt.co/product/marketplace/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=marketplace"> <img src="docs/content/imgs/readme/autogpt_marketplace.png" alt="AutoGPT Marketplace showing ready-made community agents" /> </a> <br /><strong><a href="https://agpt.co/product/marketplace/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=marketplace">Marketplace</a></strong> <br />Start from proven agents, add one to your library, and customize it for your work. </td> <td width="50%" align="center" valign="top"> <a href="https://agpt.co/product/build/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=build"> <img src="docs/content/imgs/readme/build_screen.jpg" alt="The AutoGPT Build canvas showing a real agent workflow" /> </a> <br /><strong><a href="https://agpt.co/product/build/?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=build">Build</a></strong> <br />Drag, connect, branch, and inspect blocks for exact control over every step. </td> </tr> </table> --- ## Get started ### AutoGPT Platform — public, hosted, and managed The hosted Platform is publicly available. We manage the infrastructure, model access, credentials, reliability, and updates so you can focus on the work your agents perform. **[Get started on AutoGPT Platform →](https://platform.agpt.co/signup?utm_source=github&utm_medium=referral&utm_campaign=autogpt_readme&utm_content=platform_get_started)** [Take the interactive tour →](https://platform.agpt.co/tour?utm_source=github&utm_medium=referral&utm_campaign=autogpt_readme&utm_content=tour_get_started) - AutoPilot, Agents, Marketplace, and Build - 45+ connected platforms and hundreds of AI models - No model API keys or infrastructure setup - Agents that run on demand, on schedules, and from triggers The hosted Platform is a paid service with usage-based agent runs. [Compare plans and pricing →](https://agpt.co/pricing?utm_source=github&utm_medium=referral&utm_campaign=autogpt_readme&utm_content=platform_pricing) ### Self-host AutoGPT > [!NOTE] > Self-hosting is the free path. You provide the infrastructure and model API keys, and you maintain the deployment. If you want zero setup, use the [managed Platform](https://platform.agpt.co/signup?utm_source=github&utm_medium=referral&utm_campaign=autogpt_readme&utm_content=self_host_note). **macOS and Linux:** ```bash curl -fsSL https://setup.agpt.co/install.sh -o install.sh && bash install.sh ``` **Windows PowerShell:** ```powershell powershell -c "iwr https://setup.agpt.co/install.bat -o install.bat; ./install.bat" ``` [Read the self-hosting guide →](https://docs.agpt.co/platform/getting-started) --- ## Managed Platform vs. self-hosting | | **AutoGPT Platform** | **Self-hosted** | |---|---|---| | Access | Public signup | Clone and install | | Cost | Paid plan plus agent usage | No license fee; pay your own infrastructure and model providers | | Setup | Managed | Docker and configuration required | | Model access | Built in | Bring your own API keys | | Updates and operations | Managed by AutoGPT | Managed by you | | Core builder and agent runtime | Included | Included | | Data and infrastructure control | Hosted by AutoGPT | Runs on your infrastructure | | Support | Plan-dependent | Community support | Both paths use the same repository. Choose the managed Platform when you want agents running immediately; self-host when infrastructure control matters more than operational convenience. --- ## Why the hosted Platform is paid Every agent run consumes real model usage, compute, storage, secrets management, and operational support. The managed Platform covers that infrastructure and funds continued development of the open-source project. Self-hosting remains available without a license fee for people and teams that want to provide and operate those resources themselves. --- ## What you can automate | Area | Example | |---|---| | **Executive operations** | Prepare a daily brief from internal and external signals | | **Sales** | Research every account before tomorrow's meetings | | **Marketing** | Turn a launch brief into campaign drafts across channels | | **Engineering** | Triage incidents and start with a likely cause | | **Customer support** | Draft replies, collect context, and flag escalations | | **Research** | Monitor sources and return structured reports when something changes | --- ## Integrations Connect the apps that are yours. AutoGPT provides access to hundreds of AI models and connects agents to 45+ platforms, including: `Gmail` · `Google Calendar` · `Google Docs` · `Google Sheets` · `GitHub` · `Slack` · `Discord` · `Notion` · `HubSpot` · `Linear` · `Airtable` · `Jira` · `Salesforce` · `Stripe` · `Webflow` [Explore the integrations →](https://agpt.co/docs/integrations) --- ## Community and support | Resource | Link | |---|---| | Discord | [Join the AutoGPT community](https://discord.gg/autogpt) | | Documentation | [docs.agpt.co](https://docs.agpt.co) | | Bug reports | [GitHub Issues](https://github.com/Significant-Gravitas/AutoGPT/issues/new/choose) | | Feature requests | [GitHub Discussions](https://github.com/Significant-Gravitas/AutoGPT/discussions) | | Contributing | [CONTRIBUTING.md](CONTRIBUTING.md) | --- ## License | Component | License | What it means | |---|---|---| | `autogpt_platform/` | [Polyform Shield](https://polyformproject.org/licenses/shield/1.0.0/) | Free for personal and internal business use; cannot be sold as a competing hosted service | | `classic/` and everything else | [MIT](LICENSE) | Permissive open-source use | --- ## AutoGPT Classic Looking for the original standalone AutoGPT agent? It remains available in [`classic/`](classic/) under the MIT License. - [Build an agent with Forge](classic/FORGE-QUICKSTART.md) - [Benchmark an agent with `agbenchmark`](https://pypi.org/project/agbenchmark/) - [Explore the Classic project](classic/README.md) --- ## Contributors <a href="https://github.com/Significant-Gravitas/AutoGPT/graphs/contributors"> <img src="https://contrib.rocks/image?repo=Significant-Gravitas/AutoGPT&max=1000&columns=10" alt="AutoGPT contributors" /> </a> <p align="center"> <a href="https://platform.agpt.co/signup?utm_source=github&amp;utm_medium=referral&amp;utm_campaign=autogpt_readme&amp;utm_content=footer_get_started"><strong>Get started with AutoGPT →</strong></a> </p> --- <!-- Keep these links. Translations will automatically update with the README. --> [Deutsch](https://zdoc.app/de/Significant-Gravitas/AutoGPT) | [Español](https://zdoc.app/es/Significant-Gravitas/AutoGPT) | [français](https://zdoc.app/fr/Significant-Gravitas/AutoGPT) | [日本語](https://zdoc.app/ja/Significant-Gravitas/AutoGPT) | [한국어](https://zdoc.app/ko/Significant-Gravitas/AutoGPT) | [Português](https://zdoc.app/pt/Significant-Gravitas/AutoGPT) | [Русский](https://zdoc.app/ru/Significant-Gravitas/AutoGPT) | [中文](https://zdoc.app/zh/Significant-Gravitas/AutoGPT)