# \#agents **URL:** https://community.openai.com/tag/agents/343.md [Latest](https://community.openai.com/latest.md) · [Categories](https://community.openai.com/categories.md) · [Tags](https://community.openai.com/tags.md) --- ## [Keep the Why: Git-native project memory for coding agents](https://community.openai.com/t/keep-the-why-git-native-project-memory-for-coding-agents/1404573)
**Author:** [@oliver-zehentleitner](https://community.openai.com/u/oliver-zehentleitner)\ **Replies:** 15\ **Last updated:** [October 20, 2026, 9:08pm UTC](https://community.openai.com/t/keep-the-why-git-native-project-memory-for-coding-agents/1404573 "2026-10-20T21:08:32Z")
I’ve been working on an open source project called Keep the Why. The problem I wanted to solve is pretty simple: Coding agents are getting very good at understanding code, but across sessions they often lose the reason… --- ## [AI as a tool, an automation platform, or a long-term collaborator — are these becoming different products?](https://community.openai.com/t/ai-as-a-tool-an-automation-platform-or-a-long-term-collaborator-are-these-becoming-different-products/1403250)
**Author:** [@Daareek](https://community.openai.com/u/Daareek)\ **Replies:** 12\ **Last updated:** [October 20, 2026, 7:22pm UTC](https://community.openai.com/t/ai-as-a-tool-an-automation-platform-or-a-long-term-collaborator-are-these-becoming-different-products/1403250 "2026-10-20T19:22:41Z")
I’ve been wondering whether we are starting to talk about “AI users” as if they were one group, while in reality very different usage models are emerging. The more I use AI, the more I see at least three very different … --- ## [Bug: ChatGPT Business skill sharing reports success, but recipients cannot access shared skills](https://community.openai.com/t/bug-chatgpt-business-skill-sharing-reports-success-but-recipients-cannot-access-shared-skills/1405709)
**Author:** [@Godofthemind](https://community.openai.com/u/Godofthemind)\ **Replies:** 1\ **Last updated:** [October 19, 2026, 2:47pm UTC](https://community.openai.com/t/bug-chatgpt-business-skill-sharing-reports-success-but-recipients-cannot-access-shared-skills/1405709 "2026-10-19T14:47:23Z")
Bug: ChatGPT Business skill sharing reports success, but recipients cannot access shared skills Environment Surface: ChatGPT Web Browser: Google Chrome on macOS Affected features: ChatGPT Business, Skills sharing… --- ## [Scaling Multi-Agent Coding: Keep the intelligence, Reduce the Orchestration](https://community.openai.com/t/scaling-multi-agent-coding-keep-the-intelligence-reduce-the-orchestration/1403652)
**Author:** [@maxstravion.ai](https://community.openai.com/u/maxstravion.ai)\ **Replies:** 7\ **Last updated:** [October 16, 2026, 9:09pm UTC](https://community.openai.com/t/scaling-multi-agent-coding-keep-the-intelligence-reduce-the-orchestration/1403652 "2026-10-16T21:09:11Z")
Executive summary I am currently running a large software milestone through Claude’s UltraCode orchestration. The current configuration is intentionally heavy: Opus 5.5 high reasoning effort specialist reviewer… --- ## [Workaround Request: Direct Response API for OpenAI Workspace Agents (Post-Agent Builder Deprecation)](https://community.openai.com/t/workaround-request-direct-response-api-for-openai-workspace-agents-post-agent-builder-deprecation/1393714)
**Author:** [@Vivek\_Choudhary](https://community.openai.com/u/Vivek_Choudhary)\ **Replies:** 2\ **Last updated:** [October 16, 2026, 8:17pm UTC](https://community.openai.com/t/workaround-request-direct-response-api-for-openai-workspace-agents-post-agent-builder-deprecation/1393714 "2026-10-16T20:17:28Z")
With OpenAI Agent Builder deprecating on November 30, 2026, we are looking to integrate OpenAI Workspace Agents directly into our application backend. However, we’ve run into a major architectural blocker: OpenAI’s API… --- ## [Feature Request: Make Structured Website Interfaces a Core Research Capability in ChatGPT](https://community.openai.com/t/feature-request-make-structured-website-interfaces-a-core-research-capability-in-chatgpt/1406987)
**Author:** [@scurrilus](https://community.openai.com/u/scurrilus)\ **Replies:** 0\ **Last updated:** [October 16, 2026, 3:33pm UTC](https://community.openai.com/t/feature-request-make-structured-website-interfaces-a-core-research-capability-in-chatgpt/1406987 "2026-10-16T15:33:23Z")
The idea ChatGPT should automatically discover and use structured website interfaces whenever they offer a more precise, complete or verifiable way to answer a research question than conventional web search. Websites in… --- ## [AI is creating an extraordinary divide - but not necessarily the one we think](https://community.openai.com/t/ai-is-creating-an-extraordinary-divide-but-not-necessarily-the-one-we-think/1405628)
**Author:** [@platypus](https://community.openai.com/u/platypus)\ **Replies:** 2\ **Last updated:** [October 14, 2026, 3:59pm UTC](https://community.openai.com/t/ai-is-creating-an-extraordinary-divide-but-not-necessarily-the-one-we-think/1405628 "2026-10-14T15:59:37Z")
Andrej Karpathy posted on X yesterday about the widening gap in how people experience frontier AI. Some see little more than a chatbot. A small group of professionals are watching their productivity accelerate dramatical… --- ## [Tool success is not completion evidence for an AI agent](https://community.openai.com/t/tool-success-is-not-completion-evidence-for-an-ai-agent/1405505)
**Author:** [@speralabo](https://community.openai.com/u/speralabo)\ **Replies:** 3\ **Last updated:** [October 14, 2026, 11:31am UTC](https://community.openai.com/t/tool-success-is-not-completion-evidence-for-an-ai-agent/1405505 "2026-10-14T11:31:09Z")
An agent tool call can return successfully while the intended job remains incomplete. I use three separate checks when reviewing a run: Tool result — Did the API or tool return successfully? External state — Did the t… --- ## [What finally killed “trust me” in our agent stack](https://community.openai.com/t/what-finally-killed-trust-me-in-our-agent-stack/1405460)
**Author:** [@rambozambo](https://community.openai.com/u/rambozambo)\ **Replies:** 2\ **Last updated:** [October 14, 2026, 12:41am UTC](https://community.openai.com/t/what-finally-killed-trust-me-in-our-agent-stack/1405460 "2026-10-14T00:41:55Z")
We run agents in production. For months our biggest reliability problem wasn’t the models being wrong. It was the models saying “done” when nothing happened. The patterns, in order of how often we saw them: The agent … --- ## [Does stopping a worker cancel an already queued action? A bounded QA test](https://community.openai.com/t/does-stopping-a-worker-cancel-an-already-queued-action-a-bounded-qa-test/1405895)
**Author:** [@safal0645](https://community.openai.com/u/safal0645)\ **Replies:** 2\ **Last updated:** [October 13, 2026, 8:38am UTC](https://community.openai.com/t/does-stopping-a-worker-cancel-an-already-queued-action-a-bounded-qa-test/1405895 "2026-10-13T08:38:21Z")
Hi, I’m Alexey Safonov, a QA engineer working with Codex on a small agent-action control experiment using synthetic records that I own. I’d like feedback on the cancellation boundary and the evidence needed to evaluate i… --- ## [Can Codex autonomously engineer an agent that saves everyone in RoboCup Rescue?](https://community.openai.com/t/can-codex-autonomously-engineer-an-agent-that-saves-everyone-in-robocup-rescue/1405848)
**Author:** [@Pimpcat](https://community.openai.com/u/Pimpcat)\ **Replies:** 0\ **Last updated:** [October 13, 2026, 5:22am UTC](https://community.openai.com/t/can-codex-autonomously-engineer-an-agent-that-saves-everyone-in-robocup-rescue/1405848 "2026-10-13T05:22:57Z")
I’ve been running a bit of an experiment with Codex and thought some of you might find it interesting. The project is RoboCup Rescue Agent Simulation. Basically, you’ve got ambulance teams, firefighters and police tryin… --- ## [OpenAI Agents API: reduce latency by pre-warming sandboxes](https://community.openai.com/t/openai-agents-api-reduce-latency-by-pre-warming-sandboxes/1405713)
**Author:** [@VeitB](https://community.openai.com/u/VeitB)\ **Replies:** 0\ **Last updated:** [October 12, 2026, 5:22pm UTC](https://community.openai.com/t/openai-agents-api-reduce-latency-by-pre-warming-sandboxes/1405713 "2026-10-12T17:22:46Z")
Pre-warming lets you prepare an OpenAI-hosted sandbox before creating an agent session, reducing startup latency. How it works on a high level: create an environment wait for it to be ready pass its environment\_id … --- ## [How are you deploying OpenAI Agents SDK to production?](https://community.openai.com/t/how-are-you-deploying-openai-agents-sdk-to-production/1355633)
**Author:** [@lintaixia](https://community.openai.com/u/lintaixia)\ **Replies:** 2\ **Last updated:** [October 11, 2026, 5:43pm UTC](https://community.openai.com/t/how-are-you-deploying-openai-agents-sdk-to-production/1355633 "2026-10-11T17:43:17Z")
I’m curious about how developers are deploying agents built with the OpenAI Agents SDK to production environments. My Context I’ve been experimenting with the OpenAI Agents SDK and really love how clean the API is comp… --- ## [Building Mia AI: Making AI-Agent Execution Visible, Reliable, and Safe](https://community.openai.com/t/building-mia-ai-making-ai-agent-execution-visible-reliable-and-safe/1404516)
**Author:** [@abdelshafyclapps](https://community.openai.com/u/abdelshafyclapps)\ **Replies:** 2\ **Last updated:** [October 11, 2026, 6:15am UTC](https://community.openai.com/t/building-mia-ai-making-ai-agent-execution-visible-reliable-and-safe/1404516 "2026-10-11T06:15:33Z")
Suggested tags: agents, API, tools, orchestration (select the closest available tags). Hello OpenAI Developer Community, I’m building Mia AI, a visual AI workspace designed to make working with AI agents more intuitiv… --- ## [Add a "generate password" for account CREATION in cloud browser](https://community.openai.com/t/add-a-generate-password-for-account-creation-in-cloud-browser/1404946)
**Author:** [@matthewtujague](https://community.openai.com/u/matthewtujague)\ **Replies:** 0\ **Last updated:** [October 10, 2026, 5:18am UTC](https://community.openai.com/t/add-a-generate-password-for-account-creation-in-cloud-browser/1404946 "2026-10-10T05:18:07Z")
I want my dot to apply to jobs for me, but half of these more intimate job boards require accounts, its tedious to have to whim at the computer and create one, then save it on the next go around, would be great if the ag… --- ## [Feedback: A Safety Translation and Inspection Layer for Small Vibe-Coded AI Applications](https://community.openai.com/t/feedback-a-safety-translation-and-inspection-layer-for-small-vibe-coded-ai-applications/1389448)
**Author:** [@chyperion](https://community.openai.com/u/chyperion)\ **Replies:** 1\ **Last updated:** [October 9, 2026, 5:19pm UTC](https://community.openai.com/t/feedback-a-safety-translation-and-inspection-layer-for-small-vibe-coded-ai-applications/1389448 "2026-10-09T17:19:48Z")
Feature Proposal: A Safety Translation and Inspection Layer for Small Vibe-Coded AI Applications Important Scope and Disclosure TL;DR: I have observed small vibe-coding communities informally sharing rapidly updated AI a… --- ## [Couldn't send any words at Work mode](https://community.openai.com/t/couldnt-send-any-words-at-work-mode/1403933)
**Author:** [@JuneLien](https://community.openai.com/u/JuneLien)\ **Replies:** 4\ **Last updated:** [October 7, 2026, 1:28pm UTC](https://community.openai.com/t/couldnt-send-any-words-at-work-mode/1403933 "2026-10-07T13:28:19Z")
Environment Surface: windows 11 pro 26h2 OS 26300.9550 Version : Desktop 26.1002.51308 ( just got update to this one this morning Bug What happened, and what did you expect? When I tryna use chatgpt work, no matter wh… --- ## [I built UAEP to separate agent execution success from verified reality — try to make it lie](https://community.openai.com/t/i-built-uaep-to-separate-agent-execution-success-from-verified-reality-try-to-make-it-lie/1398398)
**Author:** [@AdamMangan](https://community.openai.com/u/AdamMangan)\ **Replies:** 8\ **Last updated:** [October 5, 2026, 7:21am UTC](https://community.openai.com/t/i-built-uaep-to-separate-agent-execution-success-from-verified-reality-try-to-make-it-lie/1398398 "2026-10-05T07:21:13Z")
I’ve been working on a systems problem that I think becomes increasingly important as agents become more capable, persistent and tool-connected: what happens after the agent stack says the action succeeded? An agent ca… --- ## [Developer Mode missing and MCP app creation unavailable across multiple ChatGPT accounts](https://community.openai.com/t/developer-mode-missing-and-mcp-app-creation-unavailable-across-multiple-chatgpt-accounts/1402294)
**Author:** [@marupibneyalid](https://community.openai.com/u/marupibneyalid)\ **Replies:** 4\ **Last updated:** [October 2, 2026, 2:24pm UTC](https://community.openai.com/t/developer-mode-missing-and-mcp-app-creation-unavailable-across-multiple-chatgpt-accounts/1402294 "2026-10-02T14:24:52Z")
Hello OpenAI team and community, I would like to provide an updated and complete description of my issue. I am trying to create and use a custom MCP server with ChatGPT, but I am currently unable to complete the MCP se… --- ## [AI progress should focus on efficiency, not only more compute: lessons from CPU evolution](https://community.openai.com/t/ai-progress-should-focus-on-efficiency-not-only-more-compute-lessons-from-cpu-evolution/1402701)
**Author:** [@yewheng86](https://community.openai.com/u/yewheng86)\ **Replies:** 0\ **Last updated:** [October 1, 2026, 9:18pm UTC](https://community.openai.com/t/ai-progress-should-focus-on-efficiency-not-only-more-compute-lessons-from-cpu-evolution/1402701 "2026-10-01T21:18:03Z")
I would like to share a broader thought about how I think AI development should progress in the long term. I see an interesting parallel with the evolution of CPUs. In the earlier years of personal computing, progress … --- ## [Feature proposal: Combine context compaction, procedural memory and continual task learning for long-running agents](https://community.openai.com/t/feature-proposal-combine-context-compaction-procedural-memory-and-continual-task-learning-for-long-running-agents/1402687)
**Author:** [@yewheng86](https://community.openai.com/u/yewheng86)\ **Replies:** 0\ **Last updated:** [October 1, 2026, 8:31pm UTC](https://community.openai.com/t/feature-proposal-combine-context-compaction-procedural-memory-and-continual-task-learning-for-long-running-agents/1402687 "2026-10-01T20:31:14Z")
I recently contacted OpenAI Support with an idea about combining several AI capabilities into a single learning loop for long-running agents. Support mentioned that some of the building blocks already exist, including se… --- ## [Can I create agent that can share image on all my connected social media accounts?](https://community.openai.com/t/can-i-create-agent-that-can-share-image-on-all-my-connected-social-media-accounts/1401712)
**Author:** [@sanjayjoshi](https://community.openai.com/u/sanjayjoshi)\ **Replies:** 2\ **Last updated:** [October 12, 2026, 5:52pm UTC](https://community.openai.com/t/can-i-create-agent-that-can-share-image-on-all-my-connected-social-media-accounts/1401712 "2026-10-12T17:52:21Z")
Using buffer since 4-5 years but Now I am thinking to create social media sharing agent is okay OR continue with buffer ? --- ## [Can an autonomous agent choose silence?](https://community.openai.com/t/can-an-autonomous-agent-choose-silence/1400765)
**Author:** [@webpause22](https://community.openai.com/u/webpause22)\ **Replies:** 4\ **Last updated:** [September 27, 2026, 5:52am UTC](https://community.openai.com/t/can-an-autonomous-agent-choose-silence/1400765 "2026-09-27T05:52:34Z")
I’m testing something new with my persistent agent. It can detect real state changes, but I don’t want every change to become a notification. So I added a separate decision: Detected change ≠ Human-relevant change Th… --- ## [Give ChatGPT real agency on mobile](https://community.openai.com/t/give-chatgpt-real-agency-on-mobile/1401085)
**Author:** [@4ompik4](https://community.openai.com/u/4ompik4)\ **Replies:** 0\ **Last updated:** [September 26, 2026, 7:09pm UTC](https://community.openai.com/t/give-chatgpt-real-agency-on-mobile/1401085 "2026-09-26T19:09:28Z")
This is not just a request for app control. I think this could become one of the core directions for ChatGPT as a personal agent. I think one of the biggest next steps for ChatGPT is giving the agent real access to apps… --- ## [Is AI creating a new knowledge silo inside companies?](https://community.openai.com/t/is-ai-creating-a-new-knowledge-silo-inside-companies/1400747)
**Author:** [@Yadrima](https://community.openai.com/u/Yadrima)\ **Replies:** 1\ **Last updated:** [September 25, 2026, 7:49pm UTC](https://community.openai.com/t/is-ai-creating-a-new-knowledge-silo-inside-companies/1400747 "2026-09-25T19:49:16Z")
Something I’ve been thinking about lately: Let’s say I spend half a day with an agent debugging some weird infrastructure problem. Eventually we figure it out. Three months later someone else in the company hits almost… --- ## [Add Agents inside Codex/ChatGPT Work](https://community.openai.com/t/add-agents-inside-codex-chatgpt-work/1400694)
**Author:** [@Charles\_Sears](https://community.openai.com/u/Charles_Sears)\ **Replies:** 0\ **Last updated:** [September 25, 2026, 5:02pm UTC](https://community.openai.com/t/add-agents-inside-codex-chatgpt-work/1400694 "2026-09-25T17:02:14Z")
Hello, I am trying to sell my clients on your business plan + codex, but you guys have not added Agents to the ChatGPT Work/Codex Desktop app. Please update the desktop app so we can CRUD Agents from the desktop app. Th… --- ## [Allow Agents in the Pro Plan](https://community.openai.com/t/allow-agents-in-the-pro-plan/1400693)
**Author:** [@Charles\_Sears](https://community.openai.com/u/Charles_Sears)\ **Replies:** 0\ **Last updated:** [September 25, 2026, 4:59pm UTC](https://community.openai.com/t/allow-agents-in-the-pro-plan/1400693 "2026-09-25T16:59:04Z")
Hello, I am uncertain why it was mapped out this way, but please allow Agents inside the Pro plan. We need the same ability to grant personalities to our agents and call them for specific tasks as the business plans do. … --- ## [Provider-independent identity and delegated authority for external AI agents](https://community.openai.com/t/provider-independent-identity-and-delegated-authority-for-external-ai-agents/1390478)
**Author:** [@RagnarOath-Taker](https://community.openai.com/u/RagnarOath-Taker)\ **Replies:** 3\ **Last updated:** [September 24, 2026, 2:59pm UTC](https://community.openai.com/t/provider-independent-identity-and-delegated-authority-for-external-ai-agents/1390478 "2026-09-24T14:59:31Z")
I’ve been working on a provider-independent governance boundary for external AI agents and would value technical criticism from people building with OpenAI’s agent stack. The premise is simple: Authenticating an artifi… --- ## [Interactive browser use in ChatGPT](https://community.openai.com/t/interactive-browser-use-in-chatgpt/1400401)
**Author:** [@david.basile](https://community.openai.com/u/david.basile)\ **Replies:** 0\ **Last updated:** [September 24, 2026, 1:36pm UTC](https://community.openai.com/t/interactive-browser-use-in-chatgpt/1400401 "2026-09-24T13:36:42Z")
ChatGPT should have interactive browser use capabilities for arbitrary URLs, not just normal web fetching. the model should be able to actually use the page: click, scroll, hover, type, resize the viewport, observe anim… --- ## [Where do you draw the line between a workflow and an AI agent?](https://community.openai.com/t/where-do-you-draw-the-line-between-a-workflow-and-an-ai-agent/1399555)
**Author:** [@suryacreatx](https://community.openai.com/u/suryacreatx)\ **Replies:** 13\ **Last updated:** [September 23, 2026, 11:46am UTC](https://community.openai.com/t/where-do-you-draw-the-line-between-a-workflow-and-an-ai-agent/1399555 "2026-09-23T11:46:49Z")
I’ve been playing around with tool calling and agents recently, and I keep coming back to the same question: At what point does a workflow actually need to become an agent? For example, say I have an application where … [Next page](https://community.openai.com/tag/agents/343.md?match_all_tags=true&page=1&tags%5B%5D=agents)