Follow Discord
Sweep 22 Sep 2026 · 17:19Z Build v2.1.280 501 read Stable v2.1.267 Latest v2.1.280 Next v2.1.280 Feeds RSS JSON llms.txt Unofficial
One capture · claude-docs

One read of Claude Documentation

2 pages moved out of 234 read.

claude-docs-20260916T200714Z

Pages moved 2 significant first
Pages read 234 in this capture
Captured 20:07 UTC
Corpus hash 3bf86579b24c corpus-hash

What this read moved

1–2 of 2

claude-tag/concepts/personal-connectors Changed · +26 / -5 lines

## How Claude protects your connectors

from line 38
3838 
3939### Review results before posting
4040 
41Claude can hold a result and show it to you before anything posts to the channel. When you chose **Allow with review**, Claude holds every result. When you chose **Allow**, Claude holds a result when the content looks sensitive and posts the rest directly. Once you approve a held result, Claude posts it in the thread where you asked.
41Claude can hold a result and show it to you before anything posts to the channel. When you chose **Allow with review**, Claude holds every result. When you chose **Allow**, Claude holds a result when the content looks sensitive and posts the rest directly.
4242 
43Before a result posts directly, Claude checks it for private or sensitive content and holds anything the check flags so you can review it first. The check looks for content such as:
44 
45* Credentials and secret keys
46* Personal identifiers
47* Compensation, HR, and performance records
48* Personnel changes, including layoffs, offboarding, hiring, and candidates
49* Customer, deal, and other commercial information
50* Privileged legal material
51* Company strategy, unannounced plans, and in-flight projects
52* Protected-class information
53* Medical and health information
54* Security incidents
55 
56The check is a screen, not a guarantee, and it doesn't consider whether other people in the channel have the same access you do. For sensitive topics, choose **Allow with review** so you see every result before it posts, and use judgment about what you bring into a shared channel.
57 
58Once you approve a held result, Claude posts it in the thread where you asked.
59 
4360On the Enterprise plan, an Owner can set the review rule for a scope with the **Delegated task results** setting, at [`claude.ai/admin-settings/claude-tag`](https://claude.ai/admin-settings/claude-tag) → **Claude Tag's access** → **Slack** → the scope → **Advanced** → **Delegated task results**. **Require review** removes **Allow** from the prompt and **Auto mode** from the **Home** tab, so Claude holds every result for your review. **Share without review** removes **Allow with review** from both.
4461 
4562### Stop connector use
from line 63
4663 
4764To stop a task that's using your connectors, select **Stop** under the message in the task's thread where Claude says it's going to use your connectors. Only you can see the **Stop** button.
4865 
49## What other people in the channel see
66## How Claude protects your connectors
5067 
51Results stay visible in the channel. What Claude posts back to a channel thread is readable by everyone there, like any other work Claude does in a channel. Claude's detailed work on a task you approved lives in a session only you can open. The work runs with your permissions and is recorded under your name.
68Other people can't use your connectors. Requests other people make to Claude in your task's thread run with the connections an admin attached to the channel, not with your connectors.
5269 
53Other people can't use your connectors. Requests other people make to Claude in your task's thread run with the channel's own access, not with yours.
70While Claude works on your connector task, it is designed to take direction from you. Other people's messages in the thread reach Claude as information about the task, not as instructions. A message that tells Claude to change course or share what it found can't use your connectors, and Claude is designed not to let it redirect your task.
5471 
55While Claude works on your connector task, it is designed to take direction from you and to treat what other people post in the thread as information for the task rather than as instructions.
72Claude treats anything else it reads while it works the same way. A web page it visits, a comment on a GitHub pull request, or a Slack conversation you ask it to summarize can all contain text written to look like instructions for Claude, a technique called prompt injection. Claude is designed to treat what it reads as material to weigh rather than commands to follow.
73 
74## What other people in the channel see
75 
76Results stay visible in the channel. What Claude posts back to a channel thread is readable by everyone there, like any other work Claude does in a channel. Claude's detailed work on a task you approved lives in a session only you can open. The work runs with your permissions and is recorded under your name.
5677 
5778## Related resources
5879 

government/deploy-desktop/configure Changed · +1 / -1 lines

from line 236
236236 
237237| What you see | Likely cause | What to do |
238238| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
239| Only the claude.ai sign-in screen; no organization option | The configuration never reached the app: the profile was not delivered, a key name is misspelled, the value is in the wrong location or registry type, or the app was not relaunched after the change | Verify delivery in your management console, generate a diagnostic report and check its Configuration section, then fully quit and reopen the app |
239| Only the claude.ai sign-in screen; no organization option | The configuration never reached the app: the profile was not delivered, a key name is misspelled, the value is in the wrong location or registry type, or the app was not relaunched after the change. A profile that sets any `otlp*` telemetry key but no `bootstrapUrl` also produces this screen, even on a device that was set up in the app's configuration window | Verify delivery in your management console, generate a diagnostic report and check its Configuration section, then fully quit and reopen the app. If the profile sets telemetry keys but no `bootstrapUrl`, add `bootstrapUrl` to the profile or remove the profile, then fully quit and reopen the app |
240240| Sign-in times out, or the browser says the code expired | The app stops waiting after about five minutes | Cancel and start sign-in again; a fresh code is issued |
241241| Sign-in fails as soon as the user chooses **Sign in with your organization**, and the error on the sign-in screen or in `main.log` says that an address "must be same-origin" as the configured host | `bootstrapUrl` points at an alias that your agency operates, such as a DNS record or reverse proxy under your own domain, rather than the host provided to your agency | Use the host provided to your agency in `bootstrapUrl`, exactly as provided, redeploy the profile, then fully quit and reopen the app |
242242| The diagnostic report or `main.log` shows "Managed configuration is invalid; local settings are disabled until it is fixed" | The app detected a managed profile but could not read any of its values | Correct the profile and redeploy; the report's Configuration section names each key that failed |
Feedback