Twindem Team FAQ for controlled AI delivery --:-- UTC
Twindem

Twindem FAQ

How are company AI keys and vault entries handled?

The admin control plane supports company agent configuration and provider keys. For shared AI secrets, the vault model stores opaque ciphertext and per-user wrapped keys.

The admin control plane supports company agent configuration and provider keys. Provider keys are handled as protected server-side material, and admin APIs return metadata rather than exposing raw key values.

For shared AI secrets, the vault model stores opaque ciphertext and per-user wrapped keys. Wrapping happens client-side, while the server orchestrates custody, re-wrap operations, and audit records.

Related questions

If our developers already use Codex or Claude Code, why do we need Twindem?

AI coding tools help individual developers move faster. Twindem adds the company layer: board items, phases, second-agent review, evidence, usage, cost, and audit events attached to the work.

Open this answer
What does controlled AI delivery mean in practice?

Controlled AI delivery means AI work starts from a delivery item, moves through a defined workflow, and keeps the useful record attached: runs, summaries, findings, evidence, usage, cost, comments, handoffs, and board events.

Open this answer
Why is the second agent important?

The same agent that produces the work should not be the only reviewer of that work. Twindem separates implementation and review so a second agent can challenge the output, inspect risks, open findings, and drive a fix loop before human approval.

Open this answer
Can Twindem turn architecture work into development tasks?

Yes. Architecture work in Twindem is not just a document. It can produce concrete development task proposals, with rationale, scope, dependencies, and evidence expectations, so the team can approve what should become delivery work.

Open this answer