Feishu CLI Hits 10K Stars in Just 47 Days – Why It’s Exploding

Feishu CLI was released on March 28 and instantly earned more than 1,000 stars; 47 days later it crossed the 10,000‑star threshold, a growth curve far steeper than other office‑platform CLIs launched at the same time.

The rapid adoption is driven by AI agents: billions of CLI commands already exist in large‑model training data, so agents can understand and execute them without extra learning. When an office platform exposes its functions via a CLI, agents gain the "hands" needed to manipulate the product directly.

In a typical editorial meeting, a user simply @‑mentions the assistant; the Feishu CLI‑powered agent records the discussion into a multi‑dimensional table, generates a meeting‑summary card and an angle‑analysis card, and updates the final decision—all with only two human messages while the agent invoked the CLI 21 times across messaging, tables, cloud docs, and permission modules, shrinking a half‑hour manual process to a few seconds.

The project has attracted massive community involvement: 668 forks, over 50 external contributors, and weekly code merges. Contributors range from engineers at Turkish e‑commerce firm Hepsiburada to independent developers in Vietnam, illustrating global co‑creation.

Feishu CLI’s architecture is deliberately built for agents. It consists of three layers: Shortcut – high‑level, human‑friendly commands (e.g., "help me check today’s schedule"); API Commands – a one‑to‑one mapping to Feishu OpenAPI for precise control; and Raw API – direct access to more than 2,500 endpoints, allowing agents to use un‑wrapped APIs on the fly.

On top of the three layers, the CLI provides a Skill mechanism. Twenty‑four preset Skills cover core scenarios such as messaging, document handling, tables, calendars, and approvals, and a skill‑maker framework lets teams author custom Skills that encode reusable workflows.

Real‑time event subscription via WebSocket lets agents listen to changes in messages, schedules, or approvals and react instantly, turning the agent from a passive responder into an proactive worker. Error responses include actionable hints (e.g., switch identity or add authorization), and security defaults—input‑injection protection, output sanitisation, and key‑chain credential storage—ensure safe enterprise deployment.

Compared with peer CLIs, Feishu CLI now supports 17 business domains and over 200 commands, covering messaging, documents, multi‑dimensional tables, email, calendar, tasks, approvals, knowledge base, whiteboards, slides, meetings, OKR, attendance, cloud storage, and address book. The commands are fully interoperable; for example, an agent can fetch two weeks of calendar events, tag each meeting, write the data to a table, and generate a dashboard in a single flow.

These technical strengths, combined with massive community contributions and a clear roadmap (over 100 new capabilities and 32 releases in the first 40 days), have positioned Feishu as the de‑facto workspace where AI agents can execute real‑world office tasks, making AI a genuine digital coworker rather than a mere text generator.