AI BriefingDevin Security Swarm Launched
Key point
Cognition has launched Devin Security Swarm, which applies an Agentic MapReduce architecture to improve the efficiency of security vulnerability detection.
Details
Cognition unveiled a new security tool called Devin Security Swarm. In testing, it found 36 out of 50 real GHSA vulnerabilities, reducing the cost per vulnerability by 30% compared to the most accurate existing alternative.
To achieve this, it introduced a new architecture called Agentic MapReduce. Unlike existing agent systems that struggled to reason across entire large codebases, this approach works in the following stages.
- Map: Maps relevant signals across the repository and deploys agents focused on limited shards.
- Reduce: Consolidates the results found by the agents into a single report.
- Verify: Confirms serious vulnerabilities after verifying them through actual execution in an isolated sandbox.
Devin Security Swarm supports 14 languages including Go, Rust, and Python, and can detect complex vulnerabilities such as PHP sandbox escapes or argument injection that existing tools often miss. This tool is a core component of Devin for Security, which supports the entire process from vulnerability detection to verifying runtime exploitability and generating fix PRs (Pull Requests).
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Foremerge v0.5.0 Released: Rust-Based Open-Source Tool Detects Intent Conflicts Between Parallel AI Coding Agents