The builder is not the reviewer.
An architect proposes a Blueprint change and an independent Critic reviews it. A coding agent opens a pull request and a different reviewer reads it. The split is structural: the reviewing seat is resolved before the work starts, and the platform will not route both jobs to the same agent. Where policy requires a person, the change waits for that person.
Self-approval is blocked, not discouraged
The rule sits in the control plane, not in an instruction an agent could talk its way around. When a proposal reaches review, the platform checks which agent produced it and takes that agent out of the candidate reviewers. A workflow with no independent reviewer available stops rather than continues.
The critic reads against the Blueprint
Review here is not a second opinion on wording. The Critic checks the proposal against the Blueprint it claims to satisfy, the capabilities it composes and the policies that apply to it. Findings are recorded against the change; whether the change may proceed is settled by the policy engine, not by the Critic.
Humans keep the approvals that matter
Deterministic validation runs before any model judgement, so structural faults are caught by rules rather than by opinion. Policy then evaluates what the change touches and what it therefore requires. Anything above the line you configure — the data model, money, personal data, an external contract — waits for a named human approver.
In this capability
- Independent critic per proposal
- Author excluded from review
- Validation before model judgement
- Human approval where policy requires
AI Workforce
The rest of AI Workforce
Agent Roles
Portfolio Strategist through Release Manager: every role has a written scope, defined inputs and work it is not allowed to do.
Model Routing
Choose which model does which job. Two rules hold: no model reviews its own work, and critical workflows use more than one.
Agent Permissions
An agent holds a scope, not a key to everything: what it may propose, what it may change, what it must escalate.
Decisions, Not Transcripts
The interface shows the decision, its reasoning, its impact and the options — not a scroll of agent conversation.
Stop managing prompts. Start managing software strategy.
Connect your repositories. Let Foundariq understand what you have already built. Then tell it what your portfolio should build next.
Read-only to start · Your code stays in your GitHub · [email protected]