You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
docs(blog): add How to Build an AI Agent for Slack tutorial - #8272
Adds the builder tutorial "How to Build an AI Agent for Slack: Step-by-Step Guide (2026)" (/blog/how-to-build-an-ai-agent-for-slack).
What's Included
Astro MDX post:web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx (Prettier-verified, zero em dashes, no duplicate frontmatter description).
No actionable comments were generated in the recent review.
Summary
Summary by CodeRabbit
Documentation
Added a guide to building and configuring an AI agent for Slack, including scoping a workflow, setting instructions and permissions, configuring tools and approval flows, and connecting a Slack app.
The guide covers restricting access, testing mentions, attachments, schedules, and app-event triggers, as well as privacy controls, reusing the agent on other surfaces, and handing it off.
The pull request adds a guide to building a Slack AI agent with Agenta. It covers approach selection, workflow scoping, configuration, permissions, Slack setup, testing, automation, privacy, reuse, and handoff.
Changes: Slack AI agent guide
Layer / File(s)
Summary
Plan and configure the agent web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx
The article compares three approaches to Slack agents, defines a support-triage workflow, and describes agent configuration, instructions, skills, files, and FAQs.
Set permissions and connect Slack web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx
The article documents baseline and per-tool permissions, approval handling, Slack app connection options, and access restrictions.
Test, automate, and hand off web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx
The article describes interactive and unattended testing, automation triggers, transcript privacy, reuse across surfaces, and client handoff options.
This tutorial post has open accuracy gaps. It states an outdated Slack AI purchase requirement, skips the steps for adding the tools the walkthrough depends on, and suggests a Slack reply can approve a tool set to Deny. Readers following it could fail to complete the main workflow or be misled about costs and permissions. Fix these before publishing.
Pre-merge checks: 4
Check name
Status
Explanation
Title check
The title names a docs change that adds a blog tutorial, "How to Build an AI Agent for Slack". The changeset adds the MDX post at the matching path, plus its cover and two diagrams. The title is speci…
Description check
The description states the change: it adds the tutorial "How to Build an AI Agent for Slack: Step-by-Step Guide (2026)" at a named path. It lists the MDX file, the 1600x900 hero cover, and two diagram…
Linked Issues check
Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check
Check skipped because no linked issues were found for this pull request.
Full details: Title check
Explanation
The title names a docs change that adds a blog tutorial, "How to Build an AI Agent for Slack". The changeset adds the MDX post at the matching path, plus its cover and two diagrams. The title is specific and matches the main change. The conventional-commit prefix "docs(blog)" is accurate for a content addition.
Full details: Description check
Explanation
The description states the change: it adds the tutorial "How to Build an AI Agent for Slack: Step-by-Step Guide (2026)" at a named path. It lists the MDX file, the 1600x900 hero cover, and two diagrams with their contents. It is on-topic and specific to the changeset.
Finishing Touches
Generate unit tests (beta)
Commit to this branch
Create a new PR
Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.
Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 6 remain after this review.
Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 7 remain after this review.
The reason will be displayed to describe this comment to others. Learn more.
Minor · Functional Correctness · Quick win
Replace the obsolete Slack AI add-on requirement.
Slack includes conversation summaries and huddle notes in paid plans. Advanced features depend on the plan. Slack no longer sells the add-on through its website. Replace “requires paid Slack add-on” with the applicable plan requirement so readers can compare costs correctly. (slack.com)
Tools
LanguageTool
[style] ~47-~47: The double modal “requires paid” is nonstandard (only accepted in certain dialects). Consider “to be paid”.
Context: ... | Minutes (requires paid Slack add-on) ...
Sometimes you want one specific write action (like saving a scratch note or posting to an internal `#ops-alerts` log channel) to run automatically while every other write action pauses for review.
Open `Configuration -> Tools`, click any connected-app tool or custom MCP server tool, and set **Permission** in the `TOOL DETAILS` pane:
The reason will be displayed to describe this comment to others. Learn more.
Major · Functional Correctness · Quick win
Add the app-tool connection steps before setting permissions.
For a new agent, this instruction assumes tools that the tutorial never adds. Without those tools, the Step 5 test cannot perform the HubSpot lookup, Linear search, or ticket creation.
Add the Tools > Add tool > Third-party integration steps. Explain account authorization and action selection for HubSpot, Linear, and GitHub. Include the Slack posting action required by the scheduled digest. Agenta documents tool installation separately from permission configuration. (agenta.ai)
| **Interactive Slack Mention (`@opstriage`)** | Teammate `@mentions` or DMs the bot in Slack | `Allow` (`READ-ONLY`) | Automatic thread reply | `Ask` (human clicks **Approve** in the thread) |
| **Unattended Schedule or Webhook (`Automations`)** | UTC schedule or app event in `Sidebar > Automations` | `Allow` (`READ-ONLY`) | `Allow` (so the digest posts to `#ops-alerts` without stalling) | Draft-only in Slack or `Deny` (let a human review the digest and reply `@opstriage go ahead` in the thread) |
The reason will be displayed to describe this comment to others. Learn more.
Minor · Functional Correctness · Quick win
Do not present a Slack reply as approval for a denied tool.
If the reader sets the write tool to Deny, replying @opstriage go ahead cannot authorize that tool. Deny refuses the action regardless of the baseline policy. (agenta.ai)
Separate the two paths. Keep denied actions manual, or explain how to configure an approval-gated interactive agent before requesting the write.
In the agent's top bar (between **Templates** and **Publish**), the privacy menu controls who inside Agenta can see the agent and its Slack transcripts (`Privacy and owner`):
- **`Shared with project`:** Everyone in the Agenta project can read the agent's Slack conversations in Agenta (including direct messages sent to the bot) and send replies from Agenta that post directly back into the Slack thread. Use this for shared team workflows like `#client-support`, `#ops-alerts`, or `#team-marketing`.
- **`Private`:** Everyone permitted in the Slack channel can still `@mention` and talk to the bot in Slack, but the resulting chat transcripts inside Agenta belong exclusively to the agent's owner (plus anyone the owner explicitly invites via **Share -> Can view** or **Can send**). Use this when building an executive assistant or sensitive HR/finance triage bot.
The reason will be displayed to describe this comment to others. Learn more.
Major · Security & Privacy · Quick win
Qualify the private-chat privacy claim.
The Private bullet says transcripts belong "exclusively" to the owner and invited users. The sharing guide lists two exceptions. Project members who can read traces can read traces of a private chat (docs/docs/guides/14-share-agents-and-chats.mdx:111). Project members who can view channels can read the message logs of a channel connection through the channel API, also for a private agent (:115). Slack chats with a shared agent fall under the channel exception. Readers who choose this mode for HR or finance triage may assume a confidentiality that the product does not provide.
📝 Suggested wording
--- "a/web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx"+++ "b/web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx"@@ -292,7 +292,7 @@
In the agent's top bar (between **Templates** and **Publish**), the privacy menu controls who inside Agenta can see the agent and its Slack transcripts (`Privacy and owner`):
- **`Shared with project`:** Everyone in the Agenta project can read the agent's Slack conversations in Agenta (including direct messages sent to the bot) and send replies from Agenta that post directly back into the Slack thread. Use this for shared team workflows like `#client-support`, `#ops-alerts`, or `#team-marketing`.
-- **`Private`:** Everyone permitted in the Slack channel can still `@mention` and talk to the bot in Slack, but the resulting chat transcripts inside Agenta belong exclusively to the agent's owner (plus anyone the owner explicitly invites via **Share -> Can view** or **Can send**). Use this when building an executive assistant or sensitive HR/finance triage bot.+- **`Private`:** Everyone permitted in the Slack channel can still `@mention` and talk to the bot in Slack, but the resulting chat transcripts inside Agenta belong to the agent's owner (plus anyone the owner explicitly invites via **Share -> Can view** or **Can send**). Project members with trace access, or with channel view access, can still read some private-chat data. Review the limits in the sharing guide before you use this mode for sensitive HR or finance work.
### 2. Reuse the same agent on Telegram, WhatsApp, and website embeds
Committable suggestion
IMPORTANT
Review the replacement and indentation, then test before committing.
Suggested change
-**`Private`:** Everyone permitted in the Slack channel can still `@mention` and talk to the bot in Slack, but the resulting chat transcripts inside Agenta belong exclusively to the agent's owner (plus anyone the owner explicitly invites via **Share -> Can view** or **Can send**). Use this when building an executive assistant or sensitive HR/finance triage bot.
-**`Private`:** Everyone permitted in the Slack channel can still `@mention` and talk to the bot in Slack, but the resulting chat transcripts inside Agenta belong to the agent's owner (plus anyone the owner explicitly invites via **Share -> Can view** or **Can send**). Project members with trace access, or with channel view access, can still read some private-chat data. Review the limits in the sharing guide before you use this mode for sensitive HR or finance work.
Dependent tests finished. Post /preview for manual review. Environment absence verified; retained storage requires separate provider accounting.
Post /preview as a new comment to request a one-hour preview. Repeating it while active does not extend the hour.
Preview data is disposable. Expiry is enforced by the next successful cleanup sweep, not an exact timer.
Environment deletion stops its compute. Provider-retained storage and permanent services may still be billed.
This branch has not been deployed
No deployments
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Summary
Adds the builder tutorial "How to Build an AI Agent for Slack: Step-by-Step Guide (2026)" (
/blog/how-to-build-an-ai-agent-for-slack).What's Included
web/website/src/content/posts/how-to-build-an-ai-agent-for-slack.mdx(Prettier-verified, zero em dashes, no duplicate frontmatter description).web/website/public/blog/how-to-build-an-ai-agent-for-slack/hero.webpslack-agent-architecture.webp(3-layer progressive disclosure of instructions, skills, and shared agent files + harness/Daytona sandbox + MCP tool permissions)slack-thread-approval-and-triggers.webp(Interactive Slack thread run with👀->Allow reads-> inlineApprove / Denycard ->✅vs. unattended08:30 UTCschedule/event automations)