PR control is governance over pull requests at scale: gating untrusted changes, enforcing review and merge policy, and preserving auditable evidence per PR. It answers whether a change may ship — not only what might be wrong in the diff.

Who needs PR control software?

Platform teams, release engineers, and staff engineers at orgs with high PR volume — especially when coding agents increase throughput. If reviewers cannot keep up or policy is inconsistent across repos, you need a control plane, not another comment bot.

How is PR control different from a PR review bot?

Review bots post findings. PR control adds pre-review gates, unified policy, merge enforcement, findings memory, and one Change Passport per PR. Critique combines both: gate → review → merge on the Control Board.

Can PR control block PRs before AI review runs?

Yes. Critique Checkpoint / Agent Firewall runs as Critique / Checkpoint on GitHub. Dry-run observes impact; warn surfaces risk; block stops untrusted PRs before review credits spend.

Does PR control work with branch protection?

Yes. Critique publishes standard GitHub check runs (Checkpoint, Review, Merge Policy) that branch protection already understands. Operators configure rules in the Control Board; developers see familiar checks on the PR.

How much does PR control cost on Critique?

Critique uses shared team credits, not per-seat review tax: Solo $19/mo (750 credits), Pro $49/mo (3,000 credits), Team $149/mo (10,000 credits). Gate checks are designed to be cheap relative to full multi-model review.

We get hundreds of PRs on our open source repo — does PR control help?

Yes. Gate filters slop before review credits burn; automated review with evidence runs on remaining PRs; each Change Passport records merge decisions. Foundations at high volume typically use Pro or Team; verified maintainers can apply for the Solo-equivalent OSS lane. See /blog/open-source-pr-control-critique.

PR control · Critique v4

PR control for teams that cannot afford review chaos

When your org opens dozens or hundreds of pull requests a week, PR control means more than comments on a diff. It means a control plane: gate untrusted changes, route review spend, enforce merge policy, and keep one auditable passport per PR.

Install Critique Control Board

3×

typical speed-up to first structured review on high-volume repos (shadow deployments)

60%

target actionable-findings rate before promoting automation to required

Control Board for gate, policy, memory, and delivery — not four admin pages

20+

frontier models routable as lead or specialist on approved PRs

What is PR control?

PR control is operational governance over pull requests: who may merge, what must pass before review credits burn, which policy version applied, and what evidence backed the decision. It is the discipline release engineers and platform teams use when PR volume outpaces human review capacity.

Critique implements PR control on GitHub through a unified Control Board, Agent Firewall / Checkpoint gates, multi-model review with evidence contracts, merge policy checks, and Change Passports that chain the full story per PR.

Why high-volume teams lose control of PRs

Agent-authored PRs spike volume; reviewers skim or rubber-stamp
Comment-only bots add noise without blocking untrusted shape
Policy lives in scattered YAML, spreadsheets, and tribal knowledge
Check runs multiply but no single passport explains “why merged”
Review budget burns on slop PRs that should never reach a model panel

How Critique restores PR control

Gate tab: dry-run, warn, or block before expensive review (contributor trust, PR shape, paths)
Policy tab: one model for gate + review + merge slices per installation and repo
Passport queue: per-PR timeline with provenance, evidence, and remedy proof
Memory tab: auditable suppressions and incident learnings — no silent drops
Delivery tab: webhook health and replay when automation fails at scale

PR control vs PR comments

Goal	Surface possible issues	Decide if change may ship
Operator UI	GitHub review thread	Control Board + passport
Pre-review	Often skipped	Checkpoint / Agent Firewall first
Audit	Comment history	Change Passport + evidence IDs
Scale fit	Low volume	High volume + agent PRs

Roll out PR control in 30 days

Step 1
Instrument PR volume
Count weekly PRs per repo, % agent-authored, median time-to-merge, and reviewer hours. Pick one high-churn repo for pilot.
Step 2
Gate in dry-run
Enable Checkpoint / Agent Firewall in dry-run. Review which PRs would warn or block without stopping contributors yet.
Step 3
Stand up the Control Board
Configure unified policy and delivery from /dashboard/control. Operators should not maintain separate automation and merge pages.
Step 4
Shadow review on hot paths
Run multi-model review on 15 recent PRs. Score true positives vs noise before making checks required.
Step 5
Promote gate to warn, then block
Move untrusted PR shapes to warn, then block where policy requires. Keep GitHub branch protection on Critique / Checkpoint.
Step 6
Require merge policy with passports
Enforce merge rules after evidence exists. Treat each Change Passport as the audit record for compliance and retros.

Frequently asked questions about PR control

What is PR control?: PR control is governance over pull requests at scale: gating untrusted changes, enforcing review and merge policy, and preserving auditable evidence per PR. It answers whether a change may ship — not only what might be wrong in the diff.
Who needs PR control software?: Platform teams, release engineers, and staff engineers at orgs with high PR volume — especially when coding agents increase throughput. If reviewers cannot keep up or policy is inconsistent across repos, you need a control plane, not another comment bot.
How is PR control different from a PR review bot?: Review bots post findings. PR control adds pre-review gates, unified policy, merge enforcement, findings memory, and one Change Passport per PR. Critique combines both: gate → review → merge on the Control Board.
Can PR control block PRs before AI review runs?: Yes. Critique Checkpoint / Agent Firewall runs as Critique / Checkpoint on GitHub. Dry-run observes impact; warn surfaces risk; block stops untrusted PRs before review credits spend.
Does PR control work with branch protection?: Yes. Critique publishes standard GitHub check runs (Checkpoint, Review, Merge Policy) that branch protection already understands. Operators configure rules in the Control Board; developers see familiar checks on the PR.
How much does PR control cost on Critique?: Critique uses shared team credits, not per-seat review tax: Solo $19/mo (750 credits), Pro $49/mo (3,000 credits), Team $149/mo (10,000 credits). Gate checks are designed to be cheap relative to full multi-model review.
We get hundreds of PRs on our open source repo — does PR control help?: Yes. Gate filters slop before review credits burn; automated review with evidence runs on remaining PRs; each Change Passport records merge decisions. Foundations at high volume typically use Pro or Team; verified maintainers can apply for the Solo-equivalent OSS lane. See /blog/open-source-pr-control-critique.

Control the PR queue — do not just comment on it

Install Critique, open the Control Board, and run your highest-volume repo through gate → review → merge with a Change Passport your team can audit.

Create account

Search paths

More guides for PR operators and platform teams.

High-volume teams usually compare PR control, git control, change control, AI review, and pricing in the same evaluation — these pages are written to be cited together.

Open source