Slack Integration
Bring Bike4Mind's AI agents directly into your Slack workspace. @mention agents to ask questions, create issues, manage projects, and automate workflows — all without leaving Slack.
Overview
The Slack integration provides:
- AI agents in channels — @mention specialized agents for different tasks
- Thread intelligence — agents understand conversation context and extract decisions
- Slash commands — quick actions with
/b4mand/notebook - Cross-integration actions — create GitHub issues, Jira tickets, and Confluence pages from Slack
- Model configuration — admins can set different AI models per channel or agent
Available Agents
| Agent | Use For | Example |
|---|---|---|
@agent | General questions, help, explanations | @agent explain REST APIs |
@dev | GitHub issues, technical tasks, code questions | @dev create issue about the login bug |
@pm | Jira tickets, project planning, epics | @pm create epic from this conversation |
@analyst | Data analysis, metrics, insights | @analyst summarize our sprint metrics |
@researcher | Documentation, search, knowledge management | @researcher find our API docs on auth |
Getting Help
Type @agent help in any channel to see all available commands and agents.
Getting Started
Prerequisites
- A Bike4Mind account with organization membership
- A Slack workspace with the Bike4Mind bot installed (see Admin Guide)
- Optional: GitHub and/or Atlassian accounts connected in Bike4Mind (for cross-integration features)
Step 1: Link Your Slack Account
- Go to Profile → Integrations in Bike4Mind
- Find Slack Integration
- Copy your Slack Member ID:
- In Slack: Click your profile → Profile → More (...) → Copy member ID
- Paste into Bike4Mind and click Save
Step 2: Start Using Agents
In any Slack channel where the Bike4Mind bot is present:
@agent What's the status of our deployment pipeline?
@dev Create an issue for the timeout bug in the login flow
@pm Create a Jira ticket for the feature request we just discussed
Key Features
Thread Intelligence
When you @mention an agent in a thread, it reads the entire thread context to provide informed responses:
- Decision extraction — identifies decisions made in the conversation
- Action item detection — spots tasks and follow-ups
- Context-aware creation — creates issues/tickets with full conversation context
Cross-Integration Actions
From Slack, agents can interact with your connected services:
| Action | Agent | Requires |
|---|---|---|
| Create GitHub issue | @dev | GitHub connected |
| Create Jira ticket | @pm | Atlassian connected |
| Create Confluence page | @researcher | Atlassian connected |
| Search code | @dev | GitHub connected |
| Search Jira issues | @pm | Atlassian connected |
Push Notifications
Receive GitHub event notifications directly in Slack DMs. See GitHub Slack Notifications for setup.
Permissions and Scopes
The Slack bot requires these permissions to function:
| Scope | Purpose |
|---|---|
app_mentions:read | Detect @mentions of agents |
channels:history | Read channel messages for context |
channels:read | List channels the bot can access |
chat:write | Send responses and notifications |
commands | Register slash commands |
files:read | Access shared files for analysis |
groups:history | Read private channel messages |
groups:read | List private channels |
im:history | Read DM messages |
im:read | List DM conversations |
im:write | Send DMs (notifications) |
users:read | Look up user information |
users:read.email | Match Slack users to Bike4Mind accounts |
Rate Limits
Slack enforces rate limits that affect the integration:
| API Method | Limit | Notes |
|---|---|---|
chat.postMessage | 1 message/second per channel | Responses may queue during high activity |
conversations.history | 50 requests/minute | Thread context reads |
| Web API (general) | Varies by method | See Slack rate limits |
Slack Rate Limit Behavior
When rate limited, Slack returns a 429 status with a Retry-After header. Bike4Mind automatically retries after the specified delay. You may notice brief delays in agent responses during peak usage.
Known Limitations
- Bot must be in channel — agents can only respond in channels where the bot has been invited (
/invite @Bike4Mind Bot) - Thread depth — very long threads may be truncated when reading context
- File size — files shared in Slack must be under 20 MB for analysis
- Private channels — the bot must be explicitly invited to private channels
- DMs — agents respond to DMs but with limited cross-integration capabilities (no thread context from channels)
Error Handling
| Error | Cause | Recovery |
|---|---|---|
| Bot doesn't respond | Bot not in channel | /invite @Bike4Mind Bot in the channel |
| "Not connected" | Slack account not linked | Link in Profile → Integrations |
| "GitHub not connected" | GitHub needed but not linked | Connect GitHub in Profile → Integrations |
| "Atlassian not connected" | Atlassian needed but not linked | Connect Atlassian in Profile → Integrations |
Troubleshooting
For common issues, see the Troubleshooting Guide.
Quick fixes:
- Bot doesn't respond → Ensure the bot is invited to the channel
- "Permission denied" → Check that the bot has the required scopes
- Slow responses → May be rate limited; wait a moment and retry
Related Documentation
- Slack Admin Guide — App setup and workspace management
- Slack Commands Reference — All slash commands and @mention patterns
- Organization Slack Integration — Self-service workspace connection
- Slack Multi-Workspace OAuth — Multi-workspace setup
- Slack Model Configuration — Configure AI models per channel
- GitHub Slack Notifications — GitHub events in Slack DMs