Governing AI agents at scale with BlackRock | Devin
Devin Plugins bundle skills, rules, MCP servers, runtime hooks, and subagents with Git-backed versioning and open Agent Plugins standard compliance.
- Administrators can centrally declare plugins as required, endorsed, or forbidden at the enterprise, team, repository, or user level.
- Runtime hooks enable policy enforcement, such as restricting agent access to production infrastructure and sensitive data.
- Plugins share identical context configurations across both local and cloud-based agent environments.
- Cognition developed the governance features in collaboration with BlackRock's engineering team.
Enterprise engineering teams can centrally manage, audit, and restrict AI agent capabilities across codebases using standard Git workflows.

Sources
Read this as text
Back to the AI news