Cognition Launches Devin Auto-Triage to Monitor and Fix Production Incidents

May 18, 2026

Cognition, an applied AI lab building the Devin software engineer, launched Devin Auto-Triage as an autonomous first-responder for engineering teams. Powered by the new Devin Automations framework, these long-running agents monitor triggers like Slack or GitHub to investigate incidents immediately. This moves the agent from a reactive chat interface to an "always-on" monitoring role.

Triggers: Slack, Linear, GitHub, and others
Security: Network-sandboxed environments
Promotion: $200 in credits for first automation
Core capability: Incident investigation and PR generation
Memory type: Continuously updating, context-specific

This launch signals a shift from AI as a co-pilot to AI as a teammate capable of maintenance. It mirrors recent moves like Cursor's Sentry integration and Replit's production monitoring, where agents move beyond code generation. By using context-specific memory, Devin learns how specific teams handle bugs to improve owner assignment.

You can deploy Devin Auto-Triage by connecting it to your existing codebase and observability tools. The system runs in secure, network-sandboxed environments to protect against prompt injection during investigation. Cognition is currently offering $200 in credits for teams building their first automation, with signups available through the official Devin platform.

View the full update on devin.ai

Cognition

@cognition22h ago

Introducing Devin Auto-Triage: Your AI first-responder with long-term memory. Devin can monitor incoming bugs, alerts, and incidents, investigate them, and come back with context, next steps, or a PR.

50440

View on X

Still wondering? A few quick answers below.

Devin Auto-Triage is an autonomous AI first-responder designed to monitor engineering channels for bugs, alerts, and incidents. Unlike reactive assistants, it proactively investigates issues using a team's existing codebase and observability tools. It can deduplicate related reports, provide technical context for investigations, and even generate pull requests to resolve verified software problems.

The system operates with continuously updating, context-specific memory that allows it to learn from every incident it processes. By remembering how a specific engineering team handles past bugs, Devin can more accurately assign the right owner to new issues and reference prior investigations to speed up the resolution of recurring or related technical problems.

Cognition has built the system with several security layers to protect sensitive engineering environments. Devin and its sub-agents run in secure, network-sandboxed environments that isolate execution from the rest of the system. Additionally, the platform includes specific protections designed to prevent common AI risks such as prompt injection and unauthorized data exfiltration.

Devin Automations are long-running AI agents that power the Auto-Triage system. These agents can be assigned ongoing responsibilities rather than just responding to single prompts. They are designed to watch for triggers across various platforms like Slack, GitHub, and Linear, allowing them to act on schedules or respond to real-time events autonomously.

Engineering teams can start using the tool by creating an automation on the Devin platform and connecting it to their existing stack. For a limited time, Cognition is offering two hundred dollars in credits to users who build their first automation. The tool is designed to plug directly into existing observability and incident management workflows.