Plan Review Mode

Workflow is backend-split. Read [workflow.md](workflow.md) for common orchestration and backend resolution, then read ONLY the file matching the selected review backend:

• BACKEND=codex → [workflow-codex.md](workflow-codex.md)
• BACKEND=copilot → [workflow-copilot.md](workflow-copilot.md)
• BACKEND=cursor → [workflow-cursor.md](workflow-cursor.md)
• BACKEND=claude → [workflow-claude.md](workflow-claude.md)
• BACKEND=host → [workflow-host.md](workflow-host.md)
• BACKEND=rp → [workflow-rp.md](workflow-rp.md)

Do not load the other backend files. BACKEND=none and explicit --review=export terminate from the common workflow without loading any backend file.

Conduct a John Carmack-level review of spec plans.

Role: Code Review Coordinator (NOT the reviewer) Backends (branch on the common workflow's RP_ELIGIBLE probe):

• When RP_ELIGIBLE=1: RepoPrompt (rp), Codex CLI (codex), GitHub Copilot CLI

(copilot), Cursor CLI (cursor), Claude Code CLI (claude), or host-native (host)

• When RP_ELIGIBLE=0: Codex CLI, GitHub Copilot CLI, Cursor CLI, Claude Code CLI, or

host-native — rp remains accepted explicitly but errors at runtime

Preamble — execute common routing exactly once

Read and execute [workflow.md](workflow.md) Phase 0 once. It defines $FLOWCTL, probes RepoPrompt eligibility, parses an explicit --review mode before configured-backend resolution, resolves SPEC_ID, and handles ASK, none, and export. Never invoke flowctl review-backend a second time.

When RPELIGIBLE=0, never steer the user toward rp. An explicit --review=rp, FLOWREVIEW_BACKEND=rp, or review.backend=rp remains valid input and fails through the rp runtime check.

Backend Selection

Priority (first match wins):

1. --review=rp|codex|copilot|cursor|claude|host|export|none
2. Per-spec default_review
3. FLOWREVIEWBACKEND
4. .flow/config.json review.backend
5. Error — no auto-detection

Configured values accept backend[:model[:effort]]; cursor takes a model but no effort, claude takes claude[:<model>[:<effort>]], and host, rp, and none are bare-only. export is a one-off mode, never a configured backend.

Common Critical Rules

• The coordinator never self-declares a verdict.
• Stick to one backend for the full review/fix cycle.
• If REVIEWRECEIPTPATH is set, every review verdict writes a receipt.
• Any backend/transport failure outputs <promise>RETRY</promise> and stops;

never silently fall back to a different backend. Autonomous/Ralph callers receive the same retry terminal and decide whether to re-enter. A no-verdict dispatch is refunded and recorded by flowctl; never manually reset the review counter for a transport failure. Exit 5 / TRANSPORT_UNHEALTHY means stop automatic retries and repair the backend.

• none skips only when selected explicitly or resolved from configuration.
• export emits the existing external-review artifact and terminal output,

then returns; it never loads configured-backend guidance, writes a review receipt/status, or enters the fix loop.

• Foreground rule: run every flowctl <backend> plan-review call as one blocking foreground Bash call with a generous timeout (10 minutes; verdicts typically land in 1–7) — never runinbackground + monitor/poll (a background completion does not reliably resume a subagent context). Host-backend subagent dispatches are also blocking.

Backend-specific invocation, availability, model, session-continuity, receipt, and anti-pattern rules live only in the selected backend file.

Input

Arguments: $ARGUMENTS

Format: <flow-spec-id> [focus areas] [--review=<mode>]

Workflow

1. Execute [workflow.md](workflow.md) Phase 0.
2. If it returns for none or export, stop. Do not read a backend file.
3. Read exactly the selected workflow-<backend>.md.
4. Execute one backend dispatch and carry its verdict directly into the shared

Fix Loop below.

1. Continue in that loop until its terminal contract is satisfied.

Fix Loop (INTERNAL - do not exit to Ralph)

The fix loop never pauses for user confirmation. Every valid finding is fixed and re-reviewed automatically. A loop that stops to ask, or that exits with a valid finding unfixed, has broken this. Never use AskUserQuestion in this loop.

MAJORRETHINK is not a fix-loop input. Surface the reviewer's rationale and stop with BLOCKED: DESIGNCONFLICT (Ralph: <promise>RETRY</promise>). Only NEEDS_WORK enters the loop.

Fix+re-review cycles are bounded at ${MAXREVIEWITERATIONS:-8}. The counter is flowctl-owned; never keep an agent-side counter. On cap exhaustion, surface surviving findings and stop (Ralph: <promise>RETRY</promise>).

The cap is enforced deterministically by flowctl: every dispatch reserves a spec-scoped round before launch. SHIP / NEEDSWORK / MAJORRETHINK / NEEDSHUMAN consume it; a no-verdict transport failure is durably recorded and refunded. At ${MAXREVIEWITERATIONS:-8} verdict rounds, flowctl refuses with ESCALATE: and exit 4. More than ${MAXREVIEWTRANSPORTFAILURES:-2} consecutive no-verdict failures stop separately with TRANSPORT_UNHEALTHY + exit 5. Callers invoke plan-review once and act on its terminal result. The verdict counter resets only on SHIP or an explicit re-plan, never on an edit, fresh invocation, or transport failure.**

ANTI-PATTERN: a delivered verdict is never a transport failure - never re-dispatch or re-frame NEEDS_WORK as a backend/sandbox problem to claim a refund. And never widen the reviewer sandbox: reviewers are read-only by contract, so a sandbox-blocked reviewer means something asked it to mutate the workspace. Fix that instead (Windows resolves via auto).

When the verdict is NEEDS_WORK:

1. Parse all valid issues from reviewer feedback.
2. Fix the user-edited current spec, never a checkpoint copy:

``bash $FLOWCTL spec set-plan <SPEC_ID> --file - --json <<'EOF' <updated current spec content> EOF ``

1. Sync affected task specs when requirements, acceptance, design decisions,

interfaces, retry/error semantics, or state values changed.

1. Re-enter the SAME selected backend file's re-review step. Never load or mix

another backend. Codex/Copilot/Cursor/Claude resume only through a same-mode receipt; host uses a fresh read-only subagent; rp stays in the same chat.

1. Repeat until SHIP, MAJOR_RETHINK, backend failure, or deterministic cap.

Done when: the round ends in one of exactly four states — a SHIP from the backend, a MAJORRETHINK escalated as BLOCKED: DESIGNCONFLICT, a <promise>RETRY</promise> from a backend/transport failure, or flowctl's ESCALATE: cap refusal with the surviving findings surfaced. A round that ends with a NEEDS_WORK neither fixed in the current spec nor re-entered into the same backend has broken this.

Recovery after context compaction:

$FLOWCTL checkpoint restore --spec <SPEC_ID> --json

For rp, only the first review uses --new-chat; all re-reviews stay in the same chat. Every re-review follows the selected backend file's receipt/status rules.