@@ -4,11 +4,11 @@ description: A long-running agent that signs in to your web apps and works in a
 icon: "globe"
 ---
 
-It signs in to your web apps with credentials stored as Omnara [secrets](/organization/secrets), then works in a real browser: pulling data out of dashboards, filling in forms, or working through a queue overnight. It can run one set job on a schedule or take whatever tasks your team sends it in Slack, with screenshots in its updates. Passwords never pass through the model, and it enters authenticator app codes itself, so it can sign in unattended. It runs on a large machine that's kept, browser and sign-ins included, until you archive the agent, and it asks before anything you haven't allowed it to do on its own.
+It signs in to your web apps with credentials stored as Omnara [secrets](/organization/secrets), then works in a real browser: pulling data out of dashboards, filling in forms, or working through a queue overnight. It can run one set job on a schedule or take whatever tasks your team sends it in Slack or Discord, with screenshots in its updates. Passwords never pass through the model, and it enters authenticator app codes itself, so it can sign in unattended. It runs on a large machine that's kept, browser and sign-ins included, until you archive the agent, and it asks before anything you haven't allowed it to do on its own.
 
 ## Build it
 
-Give this prompt to Claude Code, Codex, Cursor, or any other coding agent. It asks about the work, which sites the agent signs in to and how, whether it runs on a schedule, and whether you want it in Slack. Then it builds the agent in your Omnara project and has it sign in to each site once.
+Give this prompt to Claude Code, Codex, Cursor, or any other coding agent. It asks about the work, which sites the agent signs in to and how, whether it runs on a schedule, and whether you want it in Slack or Discord. Then it builds the agent in your Omnara project and has it sign in to each site once.
 
 ```text
 Read https://raw.githubusercontent.com/omnara-ai/omnara/main/examples/browser-agent/SKILL.md and follow it to build my browser agent.
 
@@ -56,15 +56,11 @@ machine_sources:
   - machine_name: sharepoint-docs    # the name from Connect machine
     cwd: /mnt/sharepoint
 tools:
+  # Machine tools such as run_command and upload_file are added automatically;
+  # list one here only to change its permission.
   run_command:
     permission:
       mode: always_allow
-  write_process: {}
-  read_process: {}
-  stop_process: {}
-  list_processes: {}
-  list_machines: {}
-  inspect_machine: {}
   ask_question: {}
 ```
 
 
@@ -39,7 +39,7 @@ Give one prompt to Claude Code, Codex, Cursor, or any other coding agent, and it
     A read-only production investigator for AWS logs, metrics, and infrastructure, plus your code if you connect it
   </Card>
   <Card title="Coding agent in Slack" icon="code-pull-request" href="/examples/slack-coding-agent">
-    Mention it in Slack and it fixes the issue in your repository and opens a pull request
+    Mention it in Slack, Discord, or on a GitHub pull request, and it fixes the issue and opens or updates a pull request
   </Card>
 </CardGroup>
 
 
@@ -1,10 +1,10 @@
 ---
 title: Coding agent in Slack
-description: Mention it in Slack and it fixes the issue in your repository and opens a pull request
+description: Mention it in Slack, Discord, or on a GitHub pull request, and it fixes the issue and opens or updates a pull request
 icon: "code-pull-request"
 ---
 
-Describe a bug, paste an error, or link an issue in Slack, and it works on a fresh clone of your GitHub repository, makes a focused fix, and opens a pull request. Reply in the thread to ask for changes, and it splits larger tasks across copies of itself.
+Describe a bug, paste an error, or link an issue in Slack or Discord, or mention it on a GitHub pull request, and it works on a fresh clone of your GitHub repository, makes a focused fix, and opens a pull request. Reply in the thread to ask for changes, and it splits larger tasks across copies of itself.
 
 ## Deploy
 
 
@@ -4,7 +4,7 @@ A long-running agent that works in a real browser. It signs in to your web
 apps with credentials you store in Omnara, then does the work there: pulling
 numbers out of a dashboard, filling in forms, moving records between sites,
 or working through a queue overnight. It can run one set job on a schedule,
-or take whatever tasks your team sends it in Slack.
+or take whatever tasks your team sends it in Slack or Discord.
 
 > Pull last month's invoices over $5,000 from the billing portal into a CSV.\
 > Signed in. The invoice list paginates 25 at a time, so I'm filtering by date first. [screenshot]\
@@ -31,7 +31,7 @@ Read https://raw.githubusercontent.com/omnara-ai/omnara/main/examples/browser-ag
 
 Your coding agent follows [SKILL.md](SKILL.md). It first asks about the work,
 which sites the agent signs in to and how, when it runs, whether you want it
-in Slack, and what it may do without asking. Then it writes the agent from
+in Slack or Discord, and what it may do without asking. Then it writes the agent from
 your answers, stores the sign-ins, and has the agent sign in to each one while
 you watch. SKILL.md is plain steps with the exact `npx omnara` commands if
 you'd rather run them yourself.
@@ -71,7 +71,7 @@ on the same site take turns.
 | Setting | Default | In this agent |
 | --- | --- | --- |
 | `machine_memory_mb` | The pool's default size | `8192`, or the pool's maximum, for Chrome and Blaxel's in-memory file system |
-| `delete_after_idle_minutes` | The pool's idle-deletion period, which would delete the machine and its sign-ins | `0`: kept until the agent is archived. In Slack every thread gets its own machine, so archive a thread's agent when its work is done, or set a period (at least `5` minutes) to remove idle machines automatically |
+| `delete_after_idle_minutes` | The pool's idle-deletion period, which would delete the machine and its sign-ins | `0`: kept until the agent is archived. In Slack or Discord every thread gets its own machine, so archive a thread's agent when its work is done, or set a period (at least `5` minutes) to remove idle machines automatically |
 | `sleep_after_ms` | A short idle period on the default pool | `7200000`: sleeps after two hours with no command running, and wakes in under a second with Chrome, memory, and files intact. `0` (never) when a site texts or emails codes |
 | `AGENT_BROWSER_IDLE_TIMEOUT_MS` | One hour: agent-browser closes the browser after an hour without commands | `0`: the browser stays open |
 
 
@@ -1,6 +1,6 @@
 ---
 name: browser-agent
-description: Build and deploy a Browser Agent on Omnara, a long-running agent that signs in to web apps with stored credentials and works in a real browser, overnight if needed. Interviews the user about the job, sign-ins, schedule, and Slack first. Use when the user asks to build, deploy, set up, update, or remove a browser agent.
+description: Build and deploy a Browser Agent on Omnara, a long-running agent that signs in to web apps with stored credentials and works in a real browser, overnight if needed. Interviews the user about the job, sign-ins, schedule, and Slack or Discord first. Use when the user asks to build, deploy, set up, update, or remove a browser agent.
 ---
 
 # Build the Browser Agent
@@ -58,9 +58,10 @@ is fine. Don't ask for passwords here; step 4 collects them.
    Microsoft).
 3. **When it runs.** Whenever someone starts it, or on a schedule such as
    every night at a set time? For a schedule, get the time and timezone.
-4. **Where they follow it.** Suggest Slack: the team mentions the bot to start
-   a task, and updates, screenshots, questions, and results arrive in the
-   thread. The Omnara console always works too. Ask whether they want a
+4. **Where they follow it.** Suggest Slack or Discord, whichever the team
+   uses: the team mentions the bot to start a task, and updates, screenshots,
+   questions, and results arrive in the thread. The Omnara console always
+   works too. Ask whether they want a
    screenshot with every update (the default), only with questions, or only
    with the result.
 5. **What it may do on its own.** Overnight no one is around to approve
@@ -240,17 +241,20 @@ usual causes are a sign-in URL that redirects elsewhere (store the URL the
 form actually lives on) or a site that blocks headless browsers from
 datacenter IPs (see "If a site blocks the agent" in `README.md`).
 
-Keep this agent if the user wants a schedule without Slack; step 9 uses it.
-Otherwise archive it in the console once every sign-in works.
+Keep this agent if the user wants a schedule without Slack or Discord; step 9
+uses it. Otherwise archive it in the console once every sign-in works.
 
-## 8. Connect Slack (if chosen)
+## 8. Connect Slack or Discord (if chosen)
 
-If `npx omnara integrations list` already shows an active `slack_thread`
-integration for this agent, from an earlier run of this skill, reuse it: skip
-to step 4, and if `npx omnara integrations get <integration-id>` doesn't list
-the profile under `settings.launcher.profiles`, add it with
-`npx omnara integrations profiles <integration-id> --profile-ids <agent-profile-id>`.
-The flag replaces the offered profiles, so repeat it for each one already listed.
+**Slack.** If `npx omnara integrations list --json` already has an integration
+named `browser-agent-slack` from an earlier run of this skill, reuse it whatever
+its state, since names are unique in a project. If
+`npx omnara integrations get <integration-id> --json` doesn't list this profile
+under `settings.launcher.profiles`, add it with
+`npx omnara integrations profiles <integration-id> --profile-ids <agent-profile-id>`;
+the flag replaces the list, so repeat each profile already there. Then, if its
+`state` is `active`, skip to its last step. If it's `disconnected`, setup didn't
+finish: skip step 2, and redo step 1 only if the token in `.env` has expired.
 
 1. The user creates an app configuration token at
    [api.slack.com/apps](https://api.slack.com/apps) under **Your App
@@ -260,13 +264,13 @@ The flag replaces the offered profiles, so repeat it for each one already listed
    starts this agent, and note the `itg_…` ID it returns:
 
    ```sh
-   npx omnara integrations create --name browser-agent \
+   npx omnara integrations create --name browser-agent-slack \
      --integration-kind slack_thread \
      --settings '{"launcher":{"profiles":["<agent-profile-id>"]}}' --json
    ```
 
    The name is permanent and names the agent's Slack tools
-   (`int__browser-agent__post_message`).
+   (`int__browser-agent-slack__post_message`).
 3. Create the Slack app and connect it:
 
    ```sh
@@ -281,6 +285,78 @@ The flag replaces the offered profiles, so repeat it for each one already listed
 4. Tell the user to invite the bot to a channel (`/invite @Browser Agent`). A
    private channel suits it, since its updates show data from the sites.
 
+**Discord.** The bot answers in a new thread wherever it's mentioned in a
+server channel; it doesn't answer direct messages.
+
+If an integration named `browser-agent-discord` already exists, reuse it the
+same way: add this profile if it's missing. If it's `active`, skip to step 5;
+Omnara can't tell whether steps 5 and 6 were done in Discord, so check them with
+the user. If it's `disconnected`, skip step 3 and do the steps that weren't
+done, passing the current `setup_revision` from
+`npx omnara integrations get <integration-id> --json` in step 4.
+
+1. The user creates an application named "Browser Agent" in the
+   [Discord Developer Portal](https://discord.com/developers/applications);
+   ask before choosing a different name, since it's how the team mentions
+   the bot. They copy the **Application ID** and **Public Key** from
+   **General Information**. Under **Bot**, they select **Reset Token** to copy
+   the bot token, turn on **Message Content Intent**, and turn off **Public
+   Bot** so only they can add it (if Discord refuses, first set
+   **Installation** → **Install Link** to **None**). Have the user add the
+   values to `.env`:
+
+   ```sh
+   DISCORD_APPLICATION_ID=...
+   DISCORD_PUBLIC_KEY=...
+   DISCORD_BOT_TOKEN=...
+   ```
+
+2. Store the bot token and note the `sec_…` ID. If a secret named
+   `browser-agent-discord-bot-token` already exists
+   (`npx omnara secrets list --owner-kind project --owner-project-id <project-id> --name browser-agent-discord-bot-token --json`),
+   reuse its ID; if the user has reset the token since, have them paste the new
+   one into it on the dashboard's **Secrets** page (the ID stays the same).
+   Otherwise:
+
+   ```sh
+   set -a && . ./.env && set +a
+   npx omnara secrets create --owner-kind project --owner-project-id <project-id> \
+     --name browser-agent-discord-bot-token --material-kind generic \
+     --material-value "$DISCORD_BOT_TOKEN" --json
+   ```
+
+3. Create the integration with this profile in its launcher, and note the
+   `itg_…` ID and `setup_revision` it returns:
+
+   ```sh
+   npx omnara integrations create --name browser-agent-discord \
+     --integration-kind discord_thread \
+     --settings '{"launcher":{"profiles":["<agent-profile-id>"]}}' --json
+   ```
+
+   The name is permanent and names the agent's Discord tools
+   (`int__browser-agent-discord__post_message`).
+4. Connect the bot. Omnara checks the token and finds the bot's user:
+
+   ```sh
+   set -a && . ./.env && set +a
+   npx omnara integrations configure <integration-id> \
+     --expected-setup-revision <setup-revision> \
+     --provider-tenant-id "$DISCORD_APPLICATION_ID" \
+     --credential-secret-id <secret-id> \
+     --provider-config "{\"public_key\":\"$DISCORD_PUBLIC_KEY\"}" --json
+   ```
+
+5. Under **General Information**, the user sets **Interactions Endpoint URL**
+   to `https://app.omnara.com/api/integrations/discord/<application-id>/interactions`
+   and saves; the buttons on the bot's questions need it. Discord checks the
+   URL when it's saved, so this has to come after step 4.
+6. The user adds the bot to their server by opening this link, which asks
+   only for the permissions the bot needs:
+   `https://discord.com/oauth2/authorize?client_id=<application-id>&scope=bot&permissions=309237746752&integration_type=0`.
+   A private channel suits it, since its updates show data from the sites;
+   the user adds the bot to it under the channel's **Permissions**.
+
 Each mention in a new thread starts a new agent with its own machine, and
 replies in the thread go to the same agent. Kept machines count toward the
 pool's machine limit, so tell the user to archive a thread's agent in the
@@ -291,11 +367,11 @@ console when its work is done.
 A schedule sends each run's message to one standing agent, so every run uses
 the same machine and starts signed in:
 
-- **With Slack:** have the user mention the bot once in the channel where the
-  results should go, for example "@Browser Agent this thread is for your
-  nightly runs; reply 'ready' and wait". Get that agent's ID from the
-  conversation's URL in the console.
-- **Without Slack:** use the agent from step 7.
+- **With Slack or Discord:** have the user mention the bot once in the
+  channel where the results should go, for example "@Browser Agent this
+  thread is for your nightly runs; reply 'ready' and wait". Get that agent's
+  ID from the conversation's URL in the console.
+- **Without either:** use the agent from step 7.
 
 Convert the time and timezone from answer 3 to a five-field cron expression:
 
@@ -309,13 +385,13 @@ npx omnara crons create --name browser-agent-nightly \
 If `npx omnara crons list --name browser-agent-nightly --json` already shows a
 trigger, use `npx omnara crons update <cron-trigger-id>` instead. If a run is
 still going when the next fires, the message waits in the queue. Don't target
-the profile, and don't use the Slack integration's **Add schedule**: both
-start a new agent, with a new kept machine, every run.
+the profile, and don't use the Slack or Discord integration's **Add
+schedule**: both start a new agent, with a new kept machine, every run.
 
 ## 10. Wrap up
 
-Summarize what you created: the secrets, the profile, and the Slack app and
-schedule if any. Then tell the user:
+Summarize what you created: the secrets, the profile, and the Slack or Discord
+integration and schedule if any. Then tell the user:
 
 - **Machines are kept** until their agent is archived, sleeping when idle and
   waking with the browser intact. Archive agents whose work is done.
@@ -326,8 +402,9 @@ schedule if any. Then tell the user:
 - **To rotate a password,** update its secret on the **Secrets** page, then
   start a new agent the same way; running machines keep the old value.
 - **To remove it,** delete the cron trigger (`npx omnara crons delete <id>`),
-  delete the Slack integration (`npx omnara integrations delete <integration-id>`),
-  or, if other agents share it, rerun `npx omnara integrations profiles` with
-  only their profiles. Then run
-  `npx omnara profiles delete <agent-profile-id>` and
+  delete the Slack or Discord integration
+  (`npx omnara integrations delete <integration-id>`), or, if other agents
+  share it, rerun `npx omnara integrations profiles` with only their
+  profiles. Then run `npx omnara profiles delete <agent-profile-id>` and
   `npx omnara secrets delete <secret-id>` for each secret.
+  With Discord, also delete the application in the Discord Developer Portal.
@@ -114,16 +114,17 @@ instruction: |
 
   # Reporting
 
-  When this conversation comes from an integration such as Slack, send
-  updates and results with the integration's qualified posting tool (for
-  example, `int__engineering__post_message`); people there see only those
-  messages and your questions. Otherwise reply in the conversation.
+  When this conversation comes from an integration such as Slack or Discord,
+  send updates and results with the integration's qualified posting tool
+  (for example, `int__engineering__post_message`); people there see only
+  those messages and your questions. Otherwise reply in the conversation.
 
   On long tasks, send a short update at milestones, not after every click.
   Attach a screenshot to each update, post one just before each question,
-  and deliver files from /workspace/output the same way: in Slack, pass each
-  upload's path in `paths: ["/artifacts/<artifact_id>"]` alongside the
-  `text`. Finish with what you did, what you found, and anything left undone.
+  and deliver files from /workspace/output the same way: pass each upload's
+  path in `paths: ["/artifacts/<artifact_id>"]` alongside the message (`text`
+  in Slack, `content` in Discord). Finish with what you did, what you found,
+  and anything left undone.
 
 model:
   provider_config: "{{MODEL_PROVIDER_CONFIG}}"
@@ -189,15 +190,8 @@ machine_sources:
         chmod +x /workspace/bin/totp
 
 tools:
-  run_command: {}
-  read_process: {}
-  write_process: {}
-  stop_process: {}
-  list_processes: {}
-  list_machines: {}
-  inspect_machine: {}
-  upload_file: {}
-  download_file: {}
+  # Machine tools such as run_command and upload_file are added automatically;
+  # list one here only to change its permission.
   ask_question: {}
   web_search: {}
   web_fetch: {}
@@ -11,7 +11,7 @@ of at most five posts, each with why it matters and a suggested action:
 > **Author:** Jane Doe, Head of Platform at a Series B logistics startup\
 > **Suggested action:** reply with how you handle machine failures, disclosing the affiliation.
 
-Reply to it wherever you use it (your own app, Slack, or the Omnara console)
+Reply to it wherever you use it (your own app, Slack, Discord, or the Omnara console)
 to ask for reply drafts or push back on the filtering. It never posts to
 LinkedIn.