By Blockchain News | Created at 2026-06-23 17:41:22 | Updated at 2026-06-23 18:54:04
1 hour ago
Iris Coleman Jun 23, 2026 17:37
Anthropic's Claude Tag adopts 'agent identity' access model to enable secure, autonomous AI collaboration across team workspaces.
Anthropic has unveiled a new access model called 'agent identity' for its Claude Tag feature, aiming to redefine how AI agents work in team-based environments. Announced on June 24, 2026, this model is designed to provide secure, role-specific permissions for AI operations in shared workspaces, enabling enhanced collaboration without compromising security.
Claude Tag allows users to invoke Anthropic’s Claude AI in shared channels on platforms like Slack and GitHub, enabling tasks like debugging, code reviews, and data queries. By adopting the agent identity model, Claude shifts from using individual user permissions to workspace-level identities, granting admins granular control over what the AI can access and execute. This adds a layer of autonomy and security to an increasingly complex, 'multiplayer' AI ecosystem.
Why Agent Identity Matters
Traditional AI assistants operate under a single-user model, using the individual’s credentials to access tools like Google Drive or GitHub. However, this approach doesn’t scale well for multi-user team settings, where tasks often span multiple tools, roles, and permissions.
In contrast, Claude’s agent identity assigns distinct permissions at the workspace or channel level rather than relying on individual user accounts. For example, in a shared Slack channel, Claude can autonomously pull data from a GitHub repository or execute a query in a data warehouse, provided the channel's profile allows it. This ensures that no single user’s credentials become a bottleneck—or a security risk.
Key Features of Agent Identity
- Centralized Permissions: Admins set baseline permissions for Claude at the workspace level, which can then be customized for individual channels.
- Context-Specific Access: Each channel has a tailored identity, ensuring that Claude’s actions are confined to predefined scopes. For instance, legal channels can’t access engineering resources.
- Auditability: Every action performed by Claude is logged both within Anthropic’s systems and in connected tools like GitHub, providing a clear audit trail.
- Security Controls: Admins can revoke Claude’s access entirely by disabling its workspace identity, eliminating the need to manage permissions across individual user accounts.
According to Anthropic, agent identity is a direct response to the growing autonomy of AI agents. Tasks that once required constant user interaction can now be scheduled and executed independently, with Claude operating as a self-contained entity rather than an extension of a single user.
Enterprise Adoption and Future Plans
Anthropic has been steadily expanding Claude’s functionality to meet enterprise needs. In November 2025, Cognizant integrated Claude into its workflows, citing its ability to accelerate AI adoption at scale. The introduction of agent identity builds on this momentum, offering enterprises a secure yet flexible framework for deploying AI in team environments.
Looking ahead, Anthropic plans to enhance Claude Tag with just-in-time credential grants and identity-aware overlays. These features will allow organizations with complex clearance structures to implement user-level checks alongside channel permissions, further tightening security.
Implications for the AI Ecosystem
The shift from single-user to team-based AI reflects broader trends in the industry. As tools like Claude Tag gain traction, the focus is increasingly on enabling AI to function autonomously within structured boundaries. This paradigm not only enhances productivity but also addresses long-standing concerns around data security and permission management in collaborative environments.
For organizations seeking to integrate AI into their workflows, agent identity offers a compelling balance between utility and control. By centralizing permissions and providing granular access controls, Anthropic is positioning Claude as a critical tool for enterprises navigating the complexities of modern AI-driven collaboration.
Image source: Shutterstock
