Anthropic’s Claude Channels
Anthropic turned Claude into always‑on, event‑driven agents with “Claude Code Channels” that let teams control agents over Telegram and Discord — and launched a Claude Marketplace for third‑party business tools without taking an initial cut. This pushes agents into developer and chat workflows while keeping security and permission boundaries tight, signaling a platform play focused on enterprise distribution and reuse. ( )
A March 20, 2026 Claude Code changelog entry added a --channels permission-relay flag that lets channel servers forward tool-approval prompts to a mobile device. (code.claude.com) Anthropic lists Claude Marketplace as a March 6, 2026 launch and states enterprises can apply existing Anthropic spending commitments to purchase partner tools through that storefront. (claude.com) Bloomberg and SiliconANGLE report Anthropic will not take a commission on Marketplace transactions, a billing design that lets committed-spend contracts be reallocated to third‑party Claude apps. (bloomberg.com) Anthropic announced a $100 million initial commitment to a Claude Partner Network for 2026 to fund partner training, certifications, technical support, and co‑marketing. (anthropic.com) Early Marketplace and partner materials name Snowflake, Harvey, and Replit among launch integrations, while channel outreach highlights system integrators including Accenture, Deloitte, Cognizant and Infosys. (theoutpost.ai) Anthropic released an Enterprise Analytics API on February 13, 2026 that provides programmatic access to usage and engagement data for Claude and Claude Code, and Claude Code now dispatches a team of agents on every PR in a research preview—both changes that affect how enterprises must instrument agent telemetry and alerting. (support.claude.com) The Claude Code changelog also records fixes for concurrent-session OAuth refresh and improvements to MCP read/search tooling on March 20, 2026—concrete platform changes that reduce session‑stability failure modes for long‑running agent workflows. (code.claude.com) Anthropic has surfaced Claude Code Security and Cowork as limited research previews for Enterprise and Team customers, with explicit human‑review and safety warnings documented for security scanning and local‑file agent workflows. (claude.com)
