[{"data":1,"prerenderedAt":606},["ShallowReactive",2],{"help-category-\u002Fagents\u002Fcontrol-who-can-use-your-agents":3,"help-article-\u002Fagents\u002Fcontrol-who-can-use-your-agents":4,"related-articles-\u002Fagents\u002Fcontrol-who-can-use-your-agents":204},[],{"id":5,"title":6,"body":7,"category":186,"description":187,"draft":188,"extension":189,"meta":190,"navigation":191,"order":192,"path":193,"relatedArticles":194,"seo":198,"slug":199,"stem":200,"updatedAt":201,"__hash__":202,"excerpt":187,"searchText":203},"help\u002Fhelp\u002Fagents\u002F06.control-who-can-use-your-agents.md","Control who can use your agents",{"type":8,"value":9,"toc":179},"minimark",[10,14,19,22,25,29,32,125,136,146,150,156,163,176],[11,12,13],"p",{},"Who can use an agent depends on where the agent lives. Local agents are private to the machine they're on. Cloud agents belong to a workspace and follow workspace roles.",[15,16,18],"h2",{"id":17},"local-agents","Local agents",[11,20,21],{},"A local agent is only accessible on the machine where it was created. If you log in from a different device, your local agents won't appear there. No other workspace member can access a local agent through MultiClaw. To protect local agent data at the OS level, secure your device with a strong login password and full-disk encryption.",[11,23,24],{},"Use local agents for personal tasks you don't need to share.",[15,26,28],{"id":27},"cloud-agents","Cloud agents",[11,30,31],{},"Cloud agents are shared across your workspace and synced to all devices. Access is controlled by workspace roles, not by individual agent settings.",[33,34,35,58],"table",{},[36,37,38],"thead",{},[39,40,41,45,49,52,55],"tr",{},[42,43,44],"th",{},"Role",[42,46,48],{"align":47},"center","Create agents",[42,50,51],{"align":47},"Edit agents",[42,53,54],{"align":47},"Delete agents",[42,56,57],{"align":47},"Chat with agents",[59,60,61,79,94,110],"tbody",{},[39,62,63,70,73,75,77],{},[64,65,66],"td",{},[67,68,69],"strong",{},"Owner",[64,71,72],{"align":47},"Yes",[64,74,72],{"align":47},[64,76,72],{"align":47},[64,78,72],{"align":47},[39,80,81,86,88,90,92],{},[64,82,83],{},[67,84,85],{},"Admin",[64,87,72],{"align":47},[64,89,72],{"align":47},[64,91,72],{"align":47},[64,93,72],{"align":47},[39,95,96,101,104,106,108],{},[64,97,98],{},[67,99,100],{},"Member",[64,102,103],{"align":47},"No",[64,105,103],{"align":47},[64,107,103],{"align":47},[64,109,72],{"align":47},[39,111,112,117,119,121,123],{},[64,113,114],{},[67,115,116],{},"Viewer",[64,118,103],{"align":47},[64,120,103],{"align":47},[64,122,103],{"align":47},[64,124,103],{"align":47},[11,126,127,128,131,132,135],{},"Workspace owners and admins manage roles in ",[67,129,130],{},"MultiClaw Cloud"," under ",[67,133,134],{},"Settings → Members",". There is no per-agent permission system — access is workspace-role-based.",[137,138,140],"callout",{"type":139},"note",[11,141,142,143,145],{},"To allow a member to create or edit agents, promote them to ",[67,144,85],{},".",[15,147,149],{"id":148},"set-guardrails-for-everyone","Set guardrails for everyone",[11,151,152,155],{},[67,153,154],{},"Guardrails"," are restrictions that apply to all agents in your workspace, regardless of who runs them. They are the workspace owner's mechanism for enforcing policy at the organization level.",[11,157,158,159,162],{},"Configure guardrails in ",[67,160,161],{},"MultiClaw Cloud → Settings → Guardrails",". Examples of what guardrails can enforce:",[164,165,166,170,173],"ul",{},[167,168,169],"li",{},"Require approval before an agent sends any external message",[167,171,172],{},"Block agents from accessing specific domains or file paths",[167,174,175],{},"Limit agents to read-only actions in certain tools",[11,177,178],{},"Guardrails cannot be overridden by individual agents or by members. If an agent's plan is blocked by a guardrail, the agent will note this in the conversation and will typically identify which rule applied. For a full audit of guardrail enforcement, review activity logs in MultiClaw Cloud.",{"title":180,"searchDepth":181,"depth":181,"links":182},"",2,[183,184,185],{"id":17,"depth":181,"text":18},{"id":27,"depth":181,"text":28},{"id":148,"depth":181,"text":149},"agents","Understand how agent access works for local and cloud agents, and how workspace guardrails limit what all agents can do.",false,"md",{},true,6,"\u002Fagents\u002Fcontrol-who-can-use-your-agents",[195,196,197],"agents\u002Fwhat-are-agents","agents\u002Flocal-agents-vs-cloud-agents","team-governance\u002Fset-up-agent-guardrails",{"title":6,"description":187},"control-who-can-use-your-agents","help\u002Fagents\u002F06.control-who-can-use-your-agents",null,"mkGaOI3456cX-1X23sTfiLlL5PtELomLKIcdA7u1X3s","Control who can use your agents Understand how agent access works for local and cloud agents, and how workspace guardrails limit what all agents can do.",[205,358,493],{"id":206,"title":207,"body":208,"category":186,"description":345,"draft":188,"extension":189,"meta":346,"navigation":191,"order":347,"path":348,"relatedArticles":349,"seo":353,"slug":354,"stem":355,"updatedAt":201,"__hash__":356,"excerpt":345,"searchText":357},"help\u002Fhelp\u002Fagents\u002F01.what-are-agents.md","What are agents?",{"type":8,"value":209,"toc":340},[210,213,216,220,244,249,253,323,326,330,337],[11,211,212],{},"An agent is an AI assistant you configure to work for you. You give it a name, a personality, an AI model, and tools — then it can chat with you, draft plans, execute tasks, and repeat workflows automatically.",[11,214,215],{},"Before executing a task, an agent proposes a step-by-step plan for your review. You approve or reject the plan before any action is taken.",[15,217,219],{"id":218},"what-agents-can-do","What agents can do",[164,221,222,225,228,231,234,237],{},[167,223,224],{},"Answer questions and hold conversations",[167,226,227],{},"Draft step-by-step plans for tasks and wait for your approval",[167,229,230],{},"Execute tasks on a cloud desktop — browsing the web, filling forms, editing files",[167,232,233],{},"Use MCP tools to connect to apps like GitHub, Slack, Google Drive, and others",[167,235,236],{},"Run recorded browser workflows and reusable skills — each run is subject to your approval before execution",[167,238,239,240,243],{},"Delegate subtasks to other agents (",[67,241,242],{},"subagents",") — delegation is included in the plan you approve before execution begins",[137,245,246],{"type":139},[11,247,248],{},"If you configure agents to communicate with third parties — for example, sending emails or submitting forms on your behalf — you are responsible for ensuring those parties are informed they may be interacting with an automated AI system, as required by applicable law in your jurisdiction.",[15,250,252],{"id":251},"local-agents-vs-cloud-agents","Local agents vs cloud agents",[33,254,255,267],{},[36,256,257],{},[39,258,259,261,264],{},[42,260],{},[42,262,263],{},"Local agent",[42,265,266],{},"Cloud agent",[59,268,269,286,299,310],{},[39,270,271,276,284],{},[64,272,273],{},[67,274,275],{},"Where stored",[64,277,278,279,283],{},"Your machine (",[280,281,282],"code",{},"~\u002F.openclaw\u002F",")",[64,285,130],{},[39,287,288,293,296],{},[64,289,290],{},[67,291,292],{},"Who can access it",[64,294,295],{},"Only you, on that device",[64,297,298],{},"Any workspace member with permission",[39,300,301,306,308],{},[64,302,303],{},[67,304,305],{},"Syncs across devices",[64,307,103],{},[64,309,72],{},[39,311,312,317,320],{},[64,313,314],{},[67,315,316],{},"Runs 24\u002F7",[64,318,319],{},"Only while your machine is on",[64,321,322],{},"Yes, on a cloud desktop",[11,324,325],{},"Use local agents for personal work you want to keep private. Use cloud agents when your team needs shared access or you want agents running continuously.",[15,327,329],{"id":328},"agent-personalities","Agent personalities",[11,331,332,333,336],{},"Every agent has a ",[67,334,335],{},"personality"," — a system prompt that controls how it thinks, responds, and behaves. A personality might say: \"You are a senior software engineer who reviews pull requests with a focus on security and clarity.\"",[11,338,339],{},"The personality is optional, but it makes a significant difference. An agent without a personality will be generic. One with a well-written personality will feel purpose-built for your workflow.",{"title":180,"searchDepth":181,"depth":181,"links":341},[342,343,344],{"id":218,"depth":181,"text":219},{"id":251,"depth":181,"text":252},{"id":328,"depth":181,"text":329},"Agents are AI assistants that plan, act, and complete tasks on your behalf — here's what they can do and how they work.",{},1,"\u002Fagents\u002Fwhat-are-agents",[350,351,352],"agents\u002Fcreate-an-agent","agents\u002Fset-up-an-agent-profile","conversations\u002Freview-plans-and-approvals",{"title":207,"description":345},"what-are-agents","help\u002Fagents\u002F01.what-are-agents","fQhtPN58rCzs7yUj7Wctzo3MqDyxvNS6F340jCdn6C8","What are agents? Agents are AI assistants that plan, act, and complete tasks on your behalf — here's what they can do and how they work.",{"id":359,"title":252,"body":360,"category":186,"description":483,"draft":188,"extension":189,"meta":484,"navigation":191,"order":485,"path":486,"relatedArticles":487,"seo":489,"slug":251,"stem":490,"updatedAt":201,"__hash__":491,"excerpt":483,"searchText":492},"help\u002Fhelp\u002Fagents\u002F04.local-agents-vs-cloud-agents.md",{"type":8,"value":361,"toc":478},[362,373,375,438,442,445,454,457,462,466,469],[11,363,364,365,368,369,372],{},"MultiClaw supports two kinds of agents: ",[67,366,367],{},"local agents"," that live on your machine, and ",[67,370,371],{},"cloud agents"," that live in MultiClaw Cloud and run on a cloud desktop. Which kind you see depends on how your app is connected.",[15,374,252],{"id":251},[33,376,377,387],{},[36,378,379],{},[39,380,381,383,385],{},[42,382],{},[42,384,18],{},[42,386,28],{},[59,388,389,401,413,425],{},[39,390,391,395,399],{},[64,392,393],{},[67,394,275],{},[64,396,278,397,283],{},[280,398,282],{},[64,400,130],{},[39,402,403,408,411],{},[64,404,405],{},[67,406,407],{},"Who can access them",[64,409,410],{},"Only you, on the machine where they were created",[64,412,298],{},[39,414,415,420,422],{},[64,416,417],{},[67,418,419],{},"When they run",[64,421,319],{},[64,423,424],{},"Continuously on a cloud desktop (subject to platform availability)",[39,426,427,432,435],{},[64,428,429],{},[67,430,431],{},"Best for",[64,433,434],{},"Personal use, privacy-sensitive work, testing",[64,436,437],{},"Teams, automation that runs continuously",[15,439,441],{"id":440},"why-your-agent-list-may-look-different","Why your agent list may look different",[11,443,444],{},"The desktop app shows a different list depending on your connection state.",[11,446,447,448,450,451,453],{},"When you're connected to a cloud desktop, you see your ",[67,449,371],{},". When you're running in local mode — no cloud desktop connected — you see your ",[67,452,367],{},". Switching modes changes the list entirely.",[11,455,456],{},"If your agent list suddenly looks empty or unfamiliar, check the status bar at the bottom of the app. It shows whether you're connected to a cloud desktop or running locally.",[137,458,459],{"type":139},[11,460,461],{},"Local and cloud agents don't automatically sync. An agent you created locally won't appear in the cloud until you export and import it.",[15,463,465],{"id":464},"moving-agents-between-local-and-cloud","Moving agents between local and cloud",[11,467,468],{},"To use a local agent on a cloud desktop — or bring a cloud agent to your local machine — export it from the source and import it at the destination.",[11,470,471,472,477],{},"See ",[473,474,476],"a",{"href":475},"\u002Fhelp\u002Fagents\u002Fexport-and-import-agents","Export and import agents"," for step-by-step instructions.",{"title":180,"searchDepth":181,"depth":181,"links":479},[480,481,482],{"id":251,"depth":181,"text":252},{"id":440,"depth":181,"text":441},{"id":464,"depth":181,"text":465},"Understand the difference between local and cloud agents, when to use each, and why your agent list sometimes looks different.",{},4,"\u002Fagents\u002Flocal-agents-vs-cloud-agents",[195,350,488],"agents\u002Fexport-and-import-agents",{"title":252,"description":483},"help\u002Fagents\u002F04.local-agents-vs-cloud-agents","Be-4vHk5EfqKg7bbsFgtIXKdQQP5r0K2nE2eMkpGgNs","Local agents vs cloud agents Understand the difference between local and cloud agents, when to use each, and why your agent list sometimes looks different.",{"id":494,"title":495,"body":496,"category":594,"description":595,"draft":188,"extension":189,"meta":596,"navigation":191,"order":597,"path":598,"relatedArticles":599,"seo":600,"slug":601,"stem":602,"updatedAt":603,"__hash__":604,"excerpt":595,"searchText":605},"help\u002Fhelp\u002Fteam-governance\u002F03.set-up-agent-guardrails.md","Set up agent guardrails",{"type":8,"value":497,"toc":586},[498,505,509,515,522,526,531,534,538,543,546,550,555,558,562,567,570,574,581],[11,499,500,501,504],{},"Guardrails are workspace-level policies that control what your agents are allowed to do. You configure them once in ",[67,502,503],{},"MultiClaw Cloud > Settings > Guardrails",", and every agent in the workspace operates within those boundaries.",[15,506,508],{"id":507},"allowed-domains","Allowed domains",[11,510,511,512,514],{},"The ",[67,513,508],{}," list restricts which websites agents can visit during browser automation. Agents are restricted to domains on this list.",[11,516,517,518,521],{},"Add each domain you want agents to access. Use the root domain (e.g., ",[280,519,520],{},"example.com",") to include all subpages. Leave the list empty to allow all domains — adding even one entry activates enforcement.",[15,523,525],{"id":524},"blocked-tools","Blocked tools",[11,527,528,530],{},[67,529,525],{}," lets you disable specific MCP tools or entire tool categories across the workspace. For example, you can disable file deletion or block all external API calls.",[11,532,533],{},"Toggle individual tools off, or use a category toggle to disable a group at once. Blocked tools are hidden from agents entirely — agents cannot request approval to use them.",[15,535,537],{"id":536},"approval-requirements","Approval requirements",[11,539,540,542],{},[67,541,537],{}," define which action types must pause for human review before the agent proceeds. Common examples include sending emails, making purchases, and writing files.",[11,544,545],{},"Enable an action type to require approval whenever any agent attempts it. The agent pauses and awaits human review before proceeding with this action type.",[15,547,549],{"id":548},"maximum-token-budget","Maximum token budget",[11,551,552,554],{},[67,553,549],{}," caps the total tokens an agent can consume per task. This prevents runaway costs from long or looping tasks.",[11,556,557],{},"Set the cap in tokens. When an agent reaches the limit mid-task, it stops and reports the budget as exhausted. You can increase the cap at any time.",[15,559,561],{"id":560},"llm-provider-restrictions","LLM provider restrictions",[11,563,564,566],{},[67,565,561],{}," limit which AI model providers agents are permitted to use. Use this to enforce a single enterprise provider or exclude providers that don't meet your data residency requirements.",[11,568,569],{},"Select the providers you want to allow. Agents that would otherwise use a restricted provider will fail to start and display a policy error.",[15,571,573],{"id":572},"per-agent-overrides","Per-agent overrides",[11,575,576,577,580],{},"Individual agents can have additional restrictions configured on top of the workspace defaults. Per-agent guardrails can only be ",[67,578,579],{},"more"," restrictive — they cannot loosen a workspace-level policy.",[137,582,583],{"type":139},[11,584,585],{},"Guardrail policies are enforced at the platform level. Test agent behaviour in a controlled environment before deploying to production workflows.",{"title":180,"searchDepth":181,"depth":181,"links":587},[588,589,590,591,592,593],{"id":507,"depth":181,"text":508},{"id":524,"depth":181,"text":525},{"id":536,"depth":181,"text":537},{"id":548,"depth":181,"text":549},{"id":560,"depth":181,"text":561},{"id":572,"depth":181,"text":573},"team-governance","Define workspace-level policies that restrict what agents can access and do.",{},3,"\u002Fteam-governance\u002Fset-up-agent-guardrails",[],{"title":495,"description":595},"set-up-agent-guardrails","help\u002Fteam-governance\u002F03.set-up-agent-guardrails","2026-03-30","DrXDPNZtsECTZGLq6ChjZGyOyOnJnO1XvP6UxDkmOOU","Set up agent guardrails Define workspace-level policies that restrict what agents can access and do.",1778463885960]