《Cursor Docs》-Organization Groups
Organization Groups are org level cohorts in Cursor Enterprise, such as Engineering, Contractors, or Pilot Users. Members can come from any team in your Organization, and a user can belong to many groups at once.
Read More《Cursor Docs》-OpenTelemetry Export Wire Reference
Companion to OpenTelemetry Export. Full wire surface: every metric, log event, attribute, enum, and presence rule.
Read More《Cursor Docs》-OpenTelemetry Export
OpenTelemetry Export streams Cursor usage data for your team to a collector you run. Cursor sends metrics (tokens, tool calls, best effort cost) and logs (API requests, errors, corrections, skills, hooks, plugins, cloud agent lifecycle events, and recorded Grok Bot actions) to one team managed destination. Export runs server side.
Read More《Cursor Docs》-Network Configuration
Cursor needs to communicate with backend services and AI providers. This documentation covers how to configure Cursor to work within your network infrastructure, including proxies, firewalls, and encryption requirements.
Read More《Cursor Docs》-Model and Integration Management
Your team can access multiple AI models and integrate Cursor with various services. This documentation covers how to control which models are available, manage MCP server trust, and set up integrations with tools like Slack, GitHub, and Linear.
Read More《Cursor Docs》-LLM Safety and Controls
AI models can behave unexpectedly. This documentation covers how to control what agents can do, set up safety guardrails, and guide LLM behavior toward desired outcomes.
Read More《Cursor Docs》-Identity and Access Management
Identity and access management controls who can use Cursor in your organization and what they can do. You'll set up authentication, automate user provisioning, and enforce policies through device management.
Read More《Cursor Docs》-Endpoint Security Configuration
Cursor loads JavaScript modules and performs file I/O during startup. Endpoint security software that intercepts file operations or injects into processes can slow startup past internal timeouts, causing features like Agent to fail. This page covers how to configure exclusions so Cursor works alongside your security stack.
Read More《Cursor Docs》-Deployment Patterns
This guide covers how to deploy the Cursor editor and CLI tools to developer machines in your organization. Most organizations deploy both the editor (for daily development work) and the CLI (for automation, CI/CD, and scripting).
Read More《Cursor Docs》-Compliance and Monitoring
Compliance requires visibility into who did what, when, and why. This documentation covers audit logs, AI code tracking, certifications, and how to meet regulatory requirements.
Read More《Cursor Docs》-Team Pools
A pool is a named routing target that connects requests with Self Hosted Machines workers. Requests wait in the pool until an available worker claims them. Create separate pools for different execution environments, such as gpu for work that needs GPUs or ios for work that needs a Mac.
Read More《Cursor Docs》-My Machines
My Machines is the personal Self Hosted Machines configuration. It lets a specific user run Cloud Agent tool calls on a machine they already use: a laptop, devbox, or remote VM. Use it when that machine is the desired execution environment for a repo.
Read More《Cursor Docs》-Integrations
Run Team Pool workers on an infrastructure partner's platform, or start from a reference template you clone and adapt. Every option runs the same worker: the Cursor CLI opens an outbound HTTPS connection to Cursor, and Cursor sends agent tool calls over that connection.
Read More《Cursor Docs》-Computer use and desktop sharing
Computer use lets an agent on a Self Hosted Machines worker click, type, take screenshots, and drive applications with a UI. Agents can also drive a browser when the worker has Chrome or Chromium installed. It works on macOS and Linux workers, for both My Machines and Team Pools. Desktop sharing lets authorized viewers watch or take control of a Linux agent desktop from Cursor. For the product behavior on managed Cloud Agents, see Capabilities.
Read More《Cursor Docs》-Choose where Cloud Agents run
Cursor hosted Cloud Agents run each agent in an isolated cloud VM with managed lifecycle, saved environments, artifact capture, and dashboard controls for secrets and network access.
Read Moredsh-restart-web: One-click safe restart of processes in the DSH Web settings panel.
# Introduction If you use DSH via the web interface, you have likely encountered moments where the page freezes or becomes unresponsive, or where you need to restart the process for a newly installed plugin or modified configuration to take effect. The conventional approach is to return to the terminal, find the process occupying port 3080 and terminate it, then manually start DSH, and finally refresh the browser page yourself. `dsh-restart-web` streamlines this workflow into a single button in the settings panel: clicking it safely restarts the entire DSH process, and the browser automatically refreshes and reconnects after approximately 5 seconds, allowing you to stay within the web interface the entire time. DSH follows the philosophy that "everything is a plugin," and such interface-level utilities are ideally suited to be implemented as plugins. The following sections introduce its features, installation methods, and working principles. # What is dsh-restart
Read More@sal7one/dsh-design-studio: Turn design drafts into first-class workflows in DSH
# Introduction When using DeepSeek Harness (DSH) to write frontend code, the design phase often exists outside the agent workflow: screenshots are pasted into chats, changes are verbally described, and the agent opens a browser to verify its edits after modifying code. Design and coding are two separate tracks; the agent cannot see the page and cannot verify what it changed. DSH's philosophy is "everything is a plugin," and this gap can be filled by a plugin. The @sal7one/dsh-design-studio introduced below, maintained by Sal7one under an MIT license, integrates real-time preview, element selection, real design agents, and visual review into the harness, making design files directly readable, writable, and verifiable for the agent. # What is this? Defined in one sentence.
Read Moredsh-plugin-loud-failure: Turn silent tool failures into loud failures
# Foreword Fellow developers running agents have likely encountered this kind of failure: a tool call with an exit code of 0 and seemingly normal output, causing the model to proceed based on this, only to reconstruct the truth—that the initial call had actually failed—after a few steps. `pandoc` might lose a glyph but only issue a warning while still generating the PDF; pipes can swallow Python tracebacks, leaving the exit code to be provided by `tail`; or chaining commands with `;` can obscure a "command not found" error amidst a stream of normal output. The exit code indicates success, the model interprets it as success, and errors propagate downstream. While reviewing session logs afterwards can certainly reveal the cause, a more opportune moment is precisely when the failure occurs. The philosophy of DSH is that everything is a plugin; a `tool` hook is left on the tool execution chain.
Read Moredsh-web-search-toggle: Add a "Search-then-Answer" Web Search toggle for DSH sessions
**Introduction:** When developing agents using DeepSeek Harness (DSH), a common issue arises: models typically do not have internet access by default, or they only invoke `web_search` when they deem it necessary. If you require the responses in a specific session to include the latest web information, you typically have to repeatedly emphasize "search before answering" in every round of prompts. DSH's philosophy is "everything is a plugin," and such behaviors can be handled by plugins. The dsh-web-search-toggle introduced below is exactly such a plugin: it adds a "Web Search" toggle switch to the chat input toolbar. Once enabled, it forces the agent to call `web_search` before answering in every round of that session, following the official DeepSeek approach.
Read Moredsh-doc-compiled-skills: Pre-compile documents into agent-executable skills
This article introduces the DSH community skill "satan9394/dsh-doc-compiled-skills". Its core approach is to pre-compile long documents into executable skills during the build phase, rather than performing RAG retrieval at runtime, thereby making agent behavior more deterministic and reducing context overhead.
Read Moredsh-task-board: Persistent task board for DeepSeek Harness Web profile
## Introduction When running agent tasks using the DeepSeek Harness (DSH) Web profile, task goals, acceptance criteria, and execution progress are often scattered across multiple Sessions. To review what was done in a round, what passed acceptance, or what was rejected and requires retry, one must flip through session records one by one. This becomes difficult to maintain as the number of tasks increases. `dsh-task-board` solves this problem: it adds a persistent task board to the Web profile, managing task cards, execution, and review in the same interface. Below is an introduction to its features and usage. ## What is it `dsh-task-board` is a community-oriented...
Read Moresatan9394/dsh-codebase-design: A DSH plugin that requires verification before installation.
Preface: The philosophy of DSH is "everything is a plugin"; capabilities are accessed via plugins, and the community directory allows browsing and searching by entry. However, the quality of information in the directory entries varies: some plugins come with complete READMEs, installation commands, and examples, while others have only a one-line description, or documentation cannot even be found. This article introduces the latter: `satan9394/dsh-codebase-design`. As of the writing of this article (2026-09-07), its GitHub README scraping has failed, and verifiable information is extremely limited. Therefore, the focus of this article is not to list features, but to explain how to verify information in this situation before deciding whether to install. What is this? Currently, only the following points can be confirmed: - Repository URL: `https://gith`
Read Moreriffkit/dsh-plugin: A single command to install Riffkit short video skills into DeepSeek Harness
# Introduction The philosophy behind DSH is that everything is a plugin. In practice, when installing a third-party skill, there are two issues to address: it must be discoverable by the harness, yet it must not modify existing skill configurations. Patches apply whole-line replacement rather than merging to target configuration lines; if we directly overwrite the line of the base provider, the originally configured skill root will be cleared. `riffkit/dsh-plugin` handles these two issues: it packages Riffkit's agent skill into a DSH bundle that can be installed via a single command, and registers its skill root using an "insert" approach without touching existing configurations. The following sections introduce its positioning, implementation, and usage. # What is this
Read Morewhat-was-that: A small octopus to help you understand the terminology in AI responses.
# Introduction When using AI assistants to write code or conduct research, their responses often contain terms like `RAG`, `LoRA`, and `MCP`. If you don't recognize them, you have to interrupt the flow: either copy them to search elsewhere, or ask follow-up questions in the chat—the former disrupts the rhythm, while the latter pollutes the context. Here is an introduction to a small plugin that automates this: "what-was-that". It adds a little octopus desktop pet to the DSH web interface, automatically scans for terms in AI replies, and explains them in plain language. # What is it? what-was-that is a DSH static client plugin developed by SH-9999 (running on the web platform), MIT licensed, and published to npm (`what-was-that@0.1...`).
Read Moredsh-model-picker: A new two-column model selection panel for the dsh web GUI
# Introduction When developing agents using the dsh web GUI, the model selector is a high-frequency interaction point. The default implementation is a single-column, long-scrolling grouped list. With many providers and models, locating a target requires a lot of scrolling, and there is no cross-provider search. DSH's philosophy is "everything is a plugin," so the model selector for Composer can also be replaced by a plugin. The dsh-model-picker introduced below accomplishes exactly this. # What is dsh-model-picker? dsh-model-picker provides an enhanced panel for the model selector in the dsh web GUI: a provider column on the left and a model list on the right. It features independent scrolling on both sides, supports cross-provider search, and replaces the default single-column, long-scrolling grouped list. Maintained by Sanqi-normal, MIT License.
Read More