Fibery MCP

Fibery is a work management workspace where each team models its own entities, such as features, bugs, candidates, or clients, instead of accepting a fixed schema. The MCP server carries that flexibility across. Because no two Fibery workspaces look alike, it starts with introspection: an agent lists the databases in the workspace and describes one to learn its fields and types before it queries or writes anything. From there it reads and writes. Queries run against the Fibery API, so an agent can filter, sort, and page through entities, follow relations between databases, and read the activity history behind a record. Write tools create a single entity or a batch of them, update field values on existing records, and work with the documents attached to entities, which is where Fibery teams keep specs, notes, and meeting write-ups. It fits product and operations teams that have already invested in a custom Fibery schema and want an assistant that respects it rather than flattening it. The introspect-then-act pattern is the detail that matters: the agent discovers field names at runtime, so adding a field or renaming a database on the Fibery side does not require touching the integration.