1# Permission modes
2
3Permission modes control when an agent needs your approval to act. Choose a mode in the message
4composer; it applies to that thread. A new thread names its mode under the composer. In a thread
5on your default mode, open the model picker and choose **Permissions** at its bottom, or press
6`mod+shift+a`. A thread on another mode shows the mode next to the model.
7
8Set the default for new threads in **Settings → Threads → New threads → Permissions**.
9Projects can override the environment default. New threads use this setting rather than the
10mode of the thread you were viewing. The initial default is **Full access**; existing threads
11and modes you choose in a draft keep their permissions.
12
13| Mode | Behavior |
14| --------------------- | ------------------------------------------------------------------------------------- |
15| **Supervised** | Requests approval for commands and file changes. |
16| **Auto-accept edits** | Approves file edits automatically; other actions can still require approval. |
17| **Auto** | Uses the provider's automatic review to approve routine actions and ask about others. |
18| **Full access** | Allows commands and edits without approval prompts. |
19
20An administrator can cap the modes available on a machine with a
21[managed policy](./managed-policy.md#permission-modes); modes above the cap show who manages them.
22
23Approve or reject requests in the conversation to let the agent continue. Permission modes do
24not prevent the agent from asking questions about the task. Subagent approval requests appear in
25the parent conversation with the subagent's name, so you can respond without switching threads.
26
27## Provider differences
28
29Providers enforce permissions differently. Some read-only actions can proceed in **Supervised**.
30**Auto** uses automatic review on Codex, Claude, Cursor, Grok, and fx; providers without an equivalent,
31including OpenCode and Antigravity, fall back to asking. On Grok, commands its review blocks come
32to you for approval. On fx, prompts it still sends also come to you.
33
34Muse Code offers only **Supervised** and **Full access**. A Muse thread already set to another mode
35runs in **Supervised**.
36
37T3 Code offers no **Auto-accept edits** or **Full access** for fx. An fx thread already set to either runs in
38**Supervised**.
39
40Grok offers no **Auto-accept edits**. A Grok thread already set to it runs in **Supervised**. Grok
41file-change approvals offer **Allow all edits this session**. Its command approvals have no
42session-wide choice, because Grok would remember that command for the whole project.
43
44ACP Registry agents run their own tools in their own mode; T3 Code answers their approval requests
45by the permission mode. See [ACP Registry permissions](./providers-acp.md#permissions-and-terminals).
46
47Claude can still ask in **Full access** when its settings disable bypass permissions or require
48approval for an action. Dangerous-removal safety checks in bypass mode reject the command with
49the original warning. Claude can use a safe rewrite when the warning supplies one; otherwise it
50continues without the removal and tells you what it skipped.
51While a Claude approval is pending, type a text reply to decline that request and give feedback
52directly to the agent that asked, including a subagent. Other pending requests stay unanswered.
53
54Antigravity can still send native approval requests in **Full access**. It only offers remembered
55approvals for actions that support them.
56
57Kiro offers only **Supervised** and **Full access**. In **Supervised** Kiro also asks you to
58review a turn's file changes. See [Kiro permissions](./providers-kiro.md#permission-modes).
59
60See the [provider guides](./install.md#providers) for setup and provider-specific limits.