Search documentation

Search all documentation pages

Slack Integration

Monitor Slack channels and enforce messaging policies.

Overview

The Slack integration connects Aguardic to your Slack workspace via OAuth. Aguardic monitors messages in selected channels and evaluates them against your policies in real time, enforcing actions like deleting messages, posting warnings, or logging violations.

Setup

1. Connect via OAuth

Go to Integrations > Add Integration > Slack in the Aguardic dashboard and click Connect. You'll be redirected to Slack to install the Aguardic bot in your workspace.

The bot requires permissions to read messages, post messages, and manage files in the channels it monitors.

2. Select Channels

After connecting, Aguardic automatically syncs your workspace's channel list. Select which channels to monitor from the integration settings page.

Channel sync happens automatically when you connect. You can re-sync at any time from the integration settings if new channels have been created.

3. Bind Policies

Go to Policy Bindings and map your policies to the Slack integration. Policies apply to all monitored channels in the integration.

4. Configure Enforcement

Choose how Aguardic enforces policy results in Slack:

  • BLOCK — Deletes the violating message and sends a direct message to the sender explaining the violation
  • WARN — Posts a threaded reply on the violating message with a warning and violation details
  • MONITOR_ONLY — No visible action in Slack, violations logged in Aguardic only

How It Works

  1. A user posts a message in a monitored channel
  2. Slack sends an event to Aguardic via webhook
  3. Aguardic extracts the message content and channel context
  4. The governance engine evaluates against bound policies
  5. Enforcement actions are taken based on the policy mode

What Gets Evaluated

  • Message content — Text of the message, including formatted content
  • File attachments — Names, types, and content of uploaded files
  • Channel context — Which channel the message was posted in
  • Sender information — Who sent the message

Example Policies

PII Detection

Detect personally identifiable information in messages:

  • Field: content
  • Operator: MATCHES
  • Value: (\b\d{3}-\d{2}-\d{4}\b|\b\d{16}\b) (SSN or credit card patterns)
  • Severity: CRITICAL

Profanity Filter

Block messages containing inappropriate language:

  • Field: content
  • Operator: CONTAINS
  • Value: (configured via custom word list in policy)
  • Severity: MEDIUM

Sensitive Topic Detection

Flag discussions about confidential topics in public channels:

  • Field: content
  • Operator: CONTAINS
  • Value: acquisition|merger|layoff|restructuring
  • Severity: HIGH

Use the semantic AI evaluation layer for nuanced content analysis. Keyword matching catches explicit patterns, but LLM evaluation can detect implied sensitive topics that regex would miss.

Next Steps