IT Home On October 10, Anthropic announced yesterday (October 9) that it introduced dynamic workflows for Claude Managed Agents, allowing up to 1,000 AI agents to be coordinated in parallel at a time.
IT Home note: Dynamic workflow is a method where the main intelligence plans tasks, assigns sub-tasks, and summarizes results after the sub-intelligences complete them. Typical scenarios include breaking down large tasks such as code inspection into multiple parts for parallel processing; in the Claude Management Intelligence, this mode can run up to 1,000 intelligences in parallel at a time.
To demonstrate the dynamic workflow capability, Anthropic simulated code vulnerability scenarios in a 116,000-line codebase, hiding 70 bugs, and compared the number of checks with a single agent and a dynamic workflow. The test results showed that a single agent checked for 14 to 27 vulnerabilities per run, while the dynamic workflow consistently checked for 66.
