Your company's knowledge, readable by every agent.
Wiki pages, decisions, findings, and SOPs behind one governed interface — the same source of truth for people and AI, with provenance attached to every answer and contradictions raised instead of buried.
Our agents don't read documents. They read Staff.
In this studio, a file is not knowledge — if it matters, it lives in Staff Knowledge. Every answer our agents give comes through the Knowledge MCP with its source attached, and when a migrated document drifts from canon, CI blocks the change. We moved our entire document history — 264 documents — so that nothing a person once knew stayed locked in a folder.
The rule we operate by, verbatim: “If knowledge matters to Tribar, it belongs in Staff Knowledge.”
One source of truth — enforced, not promised.
A wiki people trust and a memory agents can cite. The difference between “we have documentation” and “every answer has an address, a status, and a source.”
How an agent uses it.
The full loop, from a cold start to a smarter next session.
The agent starts a session
It reads its operating instructions first — the same manual every agent in your organization runs (see Agents).
It asks the Knowledge MCP
“How do we handle client data?” The query goes to the governed surface, scoped to your organization, logged end to end.
It gets canon, with metadata
The canonical SOP, the decisions behind it, and the validation status — enough to act on, and enough to know how much to trust.
It files what it learned
Findings and decisions from the work go back into Staff Knowledge — so the next agent, or the next hire, starts smarter.
Every agent you own, grounded in what you know.
Knowledge MCP is included in Staff — self-hosted or managed. Point your first agent at it and watch the answers change.