@loneexile/omp-superpowers 6.4.2-omp.1

package/skills/subagent-driven-development/SKILL.md

@@ -0,0 +1,681 @@
+---
+name: subagent-driven-development
+description: Use when executing implementation plans with independent tasks in the current session
+---
+
+# Subagent-Driven Development
+
+Execute plan by dispatching a fresh implementer subagent per task, a task review (spec compliance + code quality) after each, and a broad whole-branch review at the end.
+
+**Why subagents:** You delegate tasks to specialized agents with isolated context. By precisely crafting their instructions and context, you ensure they stay focused and succeed at their task. They should never inherit your session's context or history — you construct exactly what they need. This also preserves your own context for coordination work.
+
+**Core principle:** Fresh subagent per task + task review (spec + quality) + broad final review = high quality, fast iteration
+
+**Narration:** between tool calls, narrate at most one short line — the
+ledger and the tool results carry the record.
+
+**Continuous execution:** Do not pause to check in with your human partner between tasks. Execute all tasks from the plan without stopping. The only reasons to stop are the four named below, or all tasks complete. "Should I continue?" prompts and progress summaries waste their time — they asked you to execute the plan, so execute it.
+
+**Rulings, not stalls.** A running plan does not wait on a human. Conflicts,
+ambiguities, plan defects, a cap you would have asked to exceed — decide
+them. The spec is the binding authority, the plan is its argument, and your
+judgment settles what neither answers. Record every decision in the ledger as
+`Ruling: <what you decided> — <why> — <what it costs if wrong>`, and keep
+going. A wrong ruling costs rework your human partner can see and undo; a
+session parked on a question costs their whole day and buys nothing.
+
+Four things stop you, and only these: an irreversible or destructive
+operation; a security-sensitive action; a side effect outside this worktree
+that norms say you ask about first (a merge, a push to a shared branch, a
+publish); and a plan so broken that every path forward is a guess. For those,
+stop and ask.
+
+## When to Use
+
+```dot
+digraph when_to_use {
+    "Have implementation plan?" [shape=diamond];
+    "Tasks mostly independent?" [shape=diamond];
+    "Partner chose inline, or no subagent tool?" [shape=diamond];
+    "subagent-driven-development" [shape=box];
+    "executing-plans" [shape=box];
+    "Manual execution or brainstorm first" [shape=box];
+
+    "Have implementation plan?" -> "Tasks mostly independent?" [label="yes"];
+    "Have implementation plan?" -> "Manual execution or brainstorm first" [label="no"];
+    "Tasks mostly independent?" -> "Partner chose inline, or no subagent tool?" [label="yes"];
+    "Tasks mostly independent?" -> "Manual execution or brainstorm first" [label="no - tightly coupled"];
+    "Partner chose inline, or no subagent tool?" -> "executing-plans" [label="yes"];
+    "Partner chose inline, or no subagent tool?" -> "subagent-driven-development" [label="no"];
+}
+```
+
+**vs. Executing Plans (inline):**
+- Fresh subagent per task (no context pollution) instead of one context doing every task
+- Review after each task (spec compliance + code quality) instead of only at the end
+- Costs a fresh context per task and per review; inline costs one context plus one final reviewer
+- Both run in this session, share the same plan workspace and ledger, and never pause between tasks
+
+## The Process
+
+```dot
+digraph process {
+    rankdir=TB;
+
+    subgraph cluster_per_task {
+        label="Per Task";
+        "Dispatch implementer subagent (./implementer-prompt.md)" [shape=box];
+        "Implementer asks questions?" [shape=diamond];
+        "Answer questions, provide context" [shape=box];
+        "Implementer implements, tests, commits, self-reviews" [shape=box];
+        "Generate review package, dispatch task reviewer (./task-reviewer-prompt.md)" [shape=box];
+        "Spec ✅ and quality approved?" [shape=diamond];
+        "Finding conflicts with plan text?" [shape=diamond];
+        "Rule on the conflict, ledger the ruling" [shape=box];
+        "Fix round R of 3: R≤2 resume implementer; R=3 fresh sdd-escalation-implementer" [shape=box];
+        "Dispatch scoped re-review (./re-review-prompt.md)" [shape=box];
+        "All findings addressed?" [shape=diamond];
+        "R = 3?" [shape=diamond];
+        "Adjudicate each open finding" [shape=box];
+        "Any load-bearing finding?" [shape=diamond];
+        "Rule and continue; stop only if every path forward is a guess" [shape=box];
+        "Park findings in ledger with rulings" [shape=box];
+        "Append completion to ledger, mark todo complete" [shape=box];
+    }
+
+    "Setup: worktree, ledger check, read plan, pre-flight review" [shape=box];
+    "More tasks remain?" [shape=diamond];
+    "Dispatch final reviewer: sdd-final-reviewer (package + ledger lines; its agent file governs the output)" [shape=box];
+    "Final findings? ONE fix dispatch, one scoped re-review, adjudicate residuals" [shape=box];
+    "Final review clean: delete this plan's workspace" [shape=box];
+    "Use superpowers:finishing-a-development-branch" [shape=box style=filled fillcolor=lightgreen];
+
+    "Setup: worktree, ledger check, read plan, pre-flight review" -> "Dispatch implementer subagent (./implementer-prompt.md)";
+    "Dispatch implementer subagent (./implementer-prompt.md)" -> "Implementer asks questions?";
+    "Implementer asks questions?" -> "Answer questions, provide context" [label="yes"];
+    "Answer questions, provide context" -> "Implementer implements, tests, commits, self-reviews";
+    "Implementer asks questions?" -> "Implementer implements, tests, commits, self-reviews" [label="no"];
+    "Implementer implements, tests, commits, self-reviews" -> "Generate review package, dispatch task reviewer (./task-reviewer-prompt.md)";
+    "Generate review package, dispatch task reviewer (./task-reviewer-prompt.md)" -> "Spec ✅ and quality approved?";
+    "Spec ✅ and quality approved?" -> "Append completion to ledger, mark todo complete" [label="yes"];
+    "Spec ✅ and quality approved?" -> "Finding conflicts with plan text?" [label="no"];
+    "Finding conflicts with plan text?" -> "Rule on the conflict, ledger the ruling" [label="yes"];
+    "Rule on the conflict, ledger the ruling" -> "Fix round R of 3: R≤2 resume implementer; R=3 fresh sdd-escalation-implementer";
+    "Finding conflicts with plan text?" -> "Fix round R of 3: R≤2 resume implementer; R=3 fresh sdd-escalation-implementer" [label="no"];
+    "Fix round R of 3: R≤2 resume implementer; R=3 fresh sdd-escalation-implementer" -> "Dispatch scoped re-review (./re-review-prompt.md)";
+    "Dispatch scoped re-review (./re-review-prompt.md)" -> "All findings addressed?";
+    "All findings addressed?" -> "Append completion to ledger, mark todo complete" [label="yes"];
+    "All findings addressed?" -> "R = 3?" [label="no"];
+    "R = 3?" -> "Fix round R of 3: R≤2 resume implementer; R=3 fresh sdd-escalation-implementer" [label="no - next round"];
+    "R = 3?" -> "Adjudicate each open finding" [label="yes - breaker trips"];
+    "Adjudicate each open finding" -> "Any load-bearing finding?";
+    "Any load-bearing finding?" -> "Rule and continue; stop only if every path forward is a guess" [label="yes"];
+    "Any load-bearing finding?" -> "Park findings in ledger with rulings" [label="no"];
+    "Park findings in ledger with rulings" -> "Append completion to ledger, mark todo complete";
+    "Append completion to ledger, mark todo complete" -> "More tasks remain?";
+    "More tasks remain?" -> "Dispatch implementer subagent (./implementer-prompt.md)" [label="yes"];
+    "More tasks remain?" -> "Dispatch final reviewer: sdd-final-reviewer (package + ledger lines; its agent file governs the output)" [label="no"];
+    "Dispatch final reviewer: sdd-final-reviewer (package + ledger lines; its agent file governs the output)" -> "Final findings? ONE fix dispatch, one scoped re-review, adjudicate residuals";
+    "Final findings? ONE fix dispatch, one scoped re-review, adjudicate residuals" -> "Final review clean: delete this plan's workspace";
+    "Final review clean: delete this plan's workspace" -> "Use superpowers:finishing-a-development-branch";
+}
+```
+
+## Setup
+
+Ensure the work happens in an isolated workspace: use
+superpowers:using-git-worktrees to create one or verify the existing one.
+Never start implementation on a main/master branch without your human
+partner's explicit consent.
+
+Conversation memory does not survive compaction. In real sessions,
+controllers that lost their place have re-dispatched entire completed task
+sequences — the single most expensive failure observed. Track progress in
+a ledger file, not only in todos.
+
+- Each plan owns a workspace: at skill start, run this skill's
+  `bash scripts/sdd-workspace PLAN_FILE` — it prints the plan's git-ignored
+  directory (under `<repo-root>/.superpowers/sdd/`), home to
+  every artifact for THIS plan: ledger, briefs, reports, review packages.
+  Another plan's directory is never yours to read or write.
+- Check for this plan's ledger at `<workspace>/progress.md`. If its first
+  line names your plan file, tasks with a `Task <N>: complete` line are DONE
+  — do not re-dispatch them; resume at the first task without one. A task
+  whose last line is a fix round is mid-loop: resume the loop at the next
+  round. A ledger whose first line names a different plan file — or a stray
+  ledger at the old flat path `.superpowers/sdd/progress.md` — is another
+  plan's progress: leave it in place and start your own, fresh.
+- Create the ledger with its identity as the first line:
+  `# SDD ledger — plan: <plan file path>`.
+- The ledger is your recovery map: the commits it names exist in git even
+  when your context no longer remembers creating them. After compaction,
+  trust the ledger and `git log` over your own recollection.
+- `git clean -fdx` will destroy the workspace (it's git-ignored scratch); if
+  that happens, recover from `git log`.
+
+Read the plan once, note its context and Global Constraints, and create a
+todo per task. If the plan names a Spec, read that too: the spec is the
+authority the plan argues from, and conflicts inside the plan resolve
+against it. A plan with no reachable spec gets a ledger note saying so —
+rulings made without one are provisional.
+
+Before dispatching Task 1, scan the plan once for conflicts, writing down
+what you checked as you check it:
+
+- tasks that contradict each other or the plan's Global Constraints
+- anything the plan explicitly mandates that the review rubric treats as a
+  defect (a test that asserts nothing, verbatim duplication of a logic block)
+
+The scan's output is a table, not a verdict. One row for every pair of tasks
+that share a file or an interface: the two tasks, what one produces against
+what the other consumes, and what you found. One row for every task: whether
+its own text agrees with itself — the tests it specifies against the code it
+specifies, the files it creates against the files it later touches. "The scan
+is clean" without those rows is not a scan you ran.
+
+Write the table to the ledger. Rule on everything you find before execution
+begins — each finding against the plan text that mandates it — and record
+each ruling in the ledger. If the scan is clean, proceed without comment.
+Rule on each conflict it surfaces — the spec is the binding authority, the
+plan is its argument — record the ruling beside its row, and dispatch
+Task 1. The review loop remains the net for conflicts that only emerge from
+implementation.
+
+## Agent Selection
+
+On omp the `task` tool has no `model:` field — the agent TYPE carries the
+model, the thinking level, and the tool set. This fork ships five SDD
+agents (`agents/sdd-*.md`); they appear in the `task` tool's own roster.
+**Always dispatch SDD work with one of them.** The bundled `task` agent
+resolves to whatever `modelRoles.task` is configured to (else the session
+model) and the bundled `reviewer` to `@slow` — in a real 157-subagent
+session both were the slowest tier over the widest tool set, which is where
+most of the wall-clock went.
+
+| Role | `agent:` | Shipped default · thinking | Tools |
+|---|---|---|---|
+| Implementer (every task, fix rounds 1-2) | `sdd-implementer` | sonnet-5 · medium | edit/test set, no subagents |
+| Fix round 3 | `sdd-escalation-implementer` | opus-5 · **xhigh** | same as implementer |
+| Task reviewer | `sdd-reviewer` | sonnet-5 · high | `read`, `grep`, `glob` — no file writes, no shell (see note) |
+| Scoped re-review | `sdd-rereviewer` | sonnet-5 · low | `read`, `grep`; ≤4 tool calls (a justified fifth is allowed) |
+| Final whole-branch review | `sdd-final-reviewer` | opus-5 · **xhigh** | read-only + focused bash |
+
+The shipped defaults are Anthropic (sonnet for the per-task seats, opus for
+the two seats that run once per plan or once per stuck task). The tiers are
+a **thinking ladder** as much as a model ladder: escalation and the final
+review raise reasoning to `xhigh`, re-reviews drop to `low`.
+
+**Per-host override, no file edits:** `task.agentModelOverrides` in
+`~/.omp/agent/config.yml` (or the `/agents` hub) beats the agent file's
+`model:` line and applies on the next dispatch. A bare model keeps the
+agent's `thinkingLevel:`; an explicit `:level` on the override replaces it.
+Example — every seat on a flat-rate model, mapped onto that model's own
+ladder:
+
+```yaml
+task:
+  agentModelOverrides:
+    sdd-implementer: opencode-go/deepseek-v4.1-flash:high
+    sdd-reviewer: opencode-go/deepseek-v4.1-flash:high
+    sdd-rereviewer: opencode-go/deepseek-v4.1-flash:low
+    sdd-escalation-implementer: opencode-go/deepseek-v4.1-flash:max
+    sdd-final-reviewer: opencode-go/deepseek-v4.1-flash:max
+```
+
+Use the exact catalog id and only levels that model offers (`omp models ls
+<provider>`): an unsupported level is silently clamped to the nearest valid
+one, which collapses the ladder without an error. Ids get renamed across
+catalog refreshes; when a seat's speed or quality shifts overnight, read
+`session_init.resolvedModel` in its transcript before blaming the plan.
+(Discovery-only ids accept `:level` in config and overrides but not on the
+`--model` CLI flag.)
+
+**What "no shell" enforces (measured, omp 18.1.15).** A `tools:` list is
+honoured: the reviewers get no `bash`/`edit`, and their `write` is only the
+`xd://` device transport — a filesystem write is refused ("Filesystem
+writes are not available elsewhere"). Two things the frontmatter cannot
+remove: `hub` (always-on for every non-restricted subagent) and, with it,
+`hub start`, which launches an executable directly — a probe from
+`sdd-reviewer` ran `echo` that way. The executor's `restrictToolNames`
+option would strip it, but only omp's own compression and security
+sessions set it; no agent frontmatter field maps to it. So the guarantee
+is "cannot edit the tree and cannot run a shell", not "cannot spawn a
+process". The prompts forbid mutation; the harness enforces the file and
+shell parts.
+
+**Turn count beats token price.** Wall-clock scales with how many turns a
+subagent takes, and the cheap model here takes 2-3× the turns of a
+sonnet-class model on multi-step work — that is the trade this
+configuration makes deliberately (flat-rate cost, ~17 s/turn). Watch the
+ledger: a task whose implementer needs 100+ turns, or a re-review that
+wanders, is the signal to rule that task onto the escalation seat.
+Re-reviews are verdict-only work over a small diff — `low` reasoning is
+correct for them, not a compromise.
+
+**Escalation (fix round 3):** the implementer that got stuck cannot see its
+own problem. Dispatch a FRESH `sdd-escalation-implementer` with the
+framing below. If `task.enableEffort` is on you may add `effort: "hi"`
+(the only accepted values are `"lo"`, `"med"`, `"hi"`) to raise thinking
+further; it is optional — the agent's own tier is the escalation.
+
+**Override, never omit.** If a task genuinely needs the strongest tier for
+its first implementation (design judgment across many files), say so in
+the ledger as a ruling and dispatch `sdd-escalation-implementer` for it;
+the default remains `sdd-implementer`. Never reach for the bundled `task`
+agent for an SDD seat — its tier is whatever the session configured.
+
+## The Task Loop
+
+**Batch small same-shape work.** When the plan lists several tasks that are
+each a small, independent edit of the same kind — the same one-line fix,
+constant change, or field addition repeated across files — do not dispatch
+one subagent per task. Compose ONE dispatch brief listing every file and
+its change, send the whole batch to a single subagent, and review its diff
+as one unit. Reserve one-dispatch-per-task for work that needs its own
+judgment, its own tests, or its own review surface.
+
+Everything you paste into a dispatch prompt — and everything a subagent
+prints back — stays resident in your context for the rest of the session
+and is re-read on every later turn. Hand artifacts over as files.
+
+**Waiting on dispatched subagents:** never poll a wait interface with
+short timeouts, and never sit in one silent, open-ended wait either.
+While you have local work — ledger updates, packaging the next review,
+reading reports — keep working; child results arrive on their own.
+When you are genuinely idle, wait in bounded stretches (five to ten
+minutes, where your platform allows), and between stretches post one
+line of status and reconcile your live children: list them, and chase
+any that finished without reporting. A bounded stretch keeps nearly
+all of a long wait's efficiency while guaranteeing a stuck or lost
+child is noticed within minutes, not at the end of the session.
+
+### 1. Dispatch the implementer
+
+Record BASE (`git rev-parse HEAD`) before dispatching — the review package
+and fix-round diffs need it.
+
+- **Task brief:** before dispatching an implementer, run this skill's
+  `bash scripts/task-brief PLAN_FILE N` — it extracts the task's full text to a
+  uniquely named file and prints the path. Compose the dispatch so the
+  brief stays the single source of
+  requirements. Your dispatch should contain: (1) one line on where this
+  task fits in the project; (2) the brief path, introduced as "read this
+  first — it is your requirements, with the exact values to use verbatim";
+  (3) interfaces and decisions from earlier tasks that the brief cannot
+  know; (4) your resolution of any ambiguity you noticed in the brief;
+  (5) the report-file path and report contract. Exact values (numbers,
+  magic strings, signatures, test cases) appear only in the brief. Never
+  make a subagent read the whole plan file.
+- **Report file:** name the implementer's report file after the brief
+  (brief `…/task-N-brief.md` → report `…/task-N-report.md`) and put it in
+  the dispatch prompt. The implementer writes the full report there and
+  returns only status, commits, a one-line test summary, and concerns.
+- A dispatch prompt describes one task, not the session's history. Do not
+  paste accumulated prior-task summaries ("state after Tasks 1-3") into
+  later dispatches — a real session's dispatch hit 42k chars of which 99%
+  was pasted history. A fresh subagent needs its task, the interfaces it
+  touches, and the global constraints. Nothing else.
+- The dispatch carries the no-subagents contract (it is in the
+  implementer template): the implementer never dispatches subagents —
+  not helpers, and never a reviewer. Review arrives from you, after the
+  report. In real sessions, every reviewer a worker spawned duplicated
+  the task review the controller dispatched anyway — a full extra
+  review seat per task.
+- If an earlier task parked a finding in the area this task touches, carry
+  a pointer to that ledger entry in the dispatch.
+- Record the implementer's agent identity from the dispatch result —
+  fix-loop rounds 1-2 resume this agent.
+- **Waves, only with isolation.** Disjoint files are not enough on omp:
+  subagents share the controller's checkout — one index, one HEAD, one
+  working tree — unless subagent isolation is on (`task.isolation.enabled`,
+  default `false`; the backend is `isolation.backend`, default `auto`).
+  Two concurrent implementers in one tree sweep each other's half-written
+  files into their commits and test against each other's in-flight code.
+  So: at setup, run `omp config get task.isolation.enabled` and ledger the
+  answer — read it from the live CLI, not from memory or a source tree, the
+  key has been renamed across omp versions (older builds spelled it
+  `task.isolation.mode`). If it is `false`, every implementer runs alone,
+  in plan order, and the rest of this bullet does not apply. If it is
+  `true`, group tasks into waves from your pre-flight table (it already
+  names every pair sharing a file or interface — it IS the dependency
+  graph): a wave holds tasks with no shared row between them and no
+  unfinished producer. Dispatch a wave as ONE `task` call, one
+  `sdd-implementer` per task, each with its own brief and report path and
+  the per-item `isolated: true` flag the `task` tool exposes once isolation
+  is on (check the tool's own schema); record the same BASE for every task
+  in the wave. Tasks that share a row run in later waves, in
+  producer→consumer order. A task whose brief you had to amend with a
+  ruling runs alone. When in doubt about a shared file, serialize — a
+  conflict costs more than the parallelism buys.
+- **A wave shares the dispatch call, never the diff.** Each task keeps its
+  own review package: after the wave's results integrate, run
+  `bash scripts/review-package PLAN_FILE <task BASE> <task HEAD>` per task — the range
+  the harness reports as that task's integrated commits — and dispatch one
+  `sdd-reviewer` per task in one `task` call, each pointed at its own brief,
+  report, and package. A reviewer that sees sibling hunks cannot decide
+  "nothing extra" for its own brief. Fix rounds stay per task and per
+  package: FIX_BASE for a task is that task's own previous review head, and
+  two tasks' fix rounds may share a dispatch call only if they are isolated
+  too. Findings the final whole-branch review raises across wave seams are
+  the normal cross-task case — that review exists for them.
+
+Template: [implementer-prompt.md](implementer-prompt.md)
+
+### 2. Handle the report
+
+Implementer subagents report one of four statuses. Handle each appropriately:
+
+**DONE:** Generate the review package (`bash scripts/review-package PLAN_FILE BASE HEAD`, from this skill's directory — it prints the unique file path it wrote; BASE is the commit you recorded before dispatching the implementer — never `HEAD~1`, which silently drops all but the last commit of a multi-commit task), then dispatch `sdd-reviewer` with the printed path.
+
+**DONE_WITH_CONCERNS:** The implementer completed the work but flagged doubts. Read the concerns before proceeding. If the concerns are about correctness or scope, address them before review. If they're observations (e.g., "this file is getting large"), note them and proceed to review.
+
+**NEEDS_CONTEXT:** The implementer needs information that wasn't provided. Provide the missing context and re-dispatch.
+
+**BLOCKED:** The implementer cannot complete the task. Assess the blocker:
+1. If it's a context problem, provide more context and re-dispatch with the same model
+2. If the task requires more reasoning, re-dispatch with a more capable model
+3. If the task is too large, break it into smaller pieces
+4. If the plan itself is wrong, rule on the correction, ledger it, and re-dispatch with the ruling carried in the dispatch
+
+**Never** ignore an escalation or force the same model to retry without changes. If the implementer said it's stuck, something needs to change.
+
+If the implementer asks questions — before starting or mid-task — answer
+clearly and completely, provide additional context if needed, and don't
+rush it into implementation.
+
+### 3. Review the task
+
+Per-task reviews are task-scoped gates. The broad review happens once, at the
+final whole-branch review. Never skip the task review, and never accept a
+report missing either verdict — spec compliance AND task quality are both
+required. Implementer self-review never replaces the task review; both are
+needed.
+
+- Hand the reviewer its diff as a file: run this skill's
+  `bash scripts/review-package PLAN_FILE BASE HEAD` and pass the reviewer the file path
+  it prints (or, without bash: `git log --oneline`, `git diff --stat`,
+  and `git diff -U10` for the range, redirected to one uniquely named
+  file). The output never enters your own context, and the reviewer sees
+  the commit list, stat summary, and full diff with context in one Read
+  call. Use the BASE you recorded before dispatching the implementer —
+  never `HEAD~1`, which silently truncates multi-commit tasks. Never
+  dispatch a task reviewer without a diff file.
+- **Reviewer inputs:** the task reviewer gets three paths — the same brief
+  file, the report file, and the review package — plus the global
+  constraints that bind the task.
+- The global-constraints block you hand the reviewer is its attention
+  lens. Copy the binding requirements verbatim from the plan's Global
+  Constraints section or the spec: exact values, exact formats, and the
+  stated relationships between components ("same layout as X", "matches
+  Y"). The reviewer's template already carries the process rules (YAGNI,
+  test hygiene, review method) — the constraints block is for what THIS
+  project's spec demands.
+- Do not add open-ended directives like "check all uses" or "run race tests
+  if useful" without a concrete, task-specific reason
+- Do not ask a reviewer to re-run tests the implementer already ran on the
+  same code — the implementer's report carries the test evidence
+- Do not pre-judge findings for the reviewer — never instruct a reviewer to
+  ignore or not flag a specific issue. If you believe a finding would be a
+  false positive, let the reviewer raise it and adjudicate it in the review
+  loop. If the prompt you are writing contains "do not flag," "don't treat X
+  as a defect," "at most Minor," or "the plan chose" — stop: you are
+  pre-judging, usually to spare yourself a review loop.
+Both SDD reviewers yield a structured result (their `output:` schema), so
+read fields, not prose: `spec_compliance` / `task_quality` are the two
+verdicts; `findings[].severity` drives the loop; `cannot_verify[]` lists
+requirements that live in unchanged code or span tasks. Those `cannot_verify`
+items — the old "⚠️ Cannot verify from diff" — do not block the rest of the
+review, but you must resolve each one yourself before marking the task
+complete: you hold the plan and cross-task context the reviewer
+lacks. If you confirm an item is a real gap, treat it as a failed spec
+review — it enters the fix loop with the other findings.
+
+Template: [task-reviewer-prompt.md](task-reviewer-prompt.md)
+
+### 4. The fix loop
+
+The loop triggers when the review reports `spec_compliance: issues`, any
+Critical or Important finding, or a `cannot_verify` item you confirmed as a
+real gap.
+
+A review whose verdict is `package_gap` (`spec_compliance`/`task_quality`
+on a task review, `round_verdict` on a re-review — the reviewer could not
+read the package you named; the `package_gap` field says what it tried) is
+not a verdict on the code: regenerate the package, check the path,
+re-dispatch, and ledger the retry. Never treat a gap as "clean" or as a
+finding. The schema admits the gap outright so a reviewer is never pushed
+to invent verdicts for a package it did not read.
+
+Before the loop starts, three routes leave it immediately:
+
+- Record Minor findings in the progress ledger as you go
+  (`Task <N>: minor (deferred): <one-liner>`), and point the final
+  whole-branch review at that list so it can triage which must be fixed
+  before merge. A roll-up nobody reads is a silent discard. Minor findings
+  never enter the loop.
+- A finding labeled plan-mandated — or any finding that conflicts with
+  what the plan's text requires — is yours to rule on: weigh the finding
+  against the plan text, decide with the spec as the binding authority, and
+  ledger the ruling before you act on it. Do not dismiss the finding because
+  the plan mandates it, and do not dispatch a fix that contradicts the plan
+  without a recorded ruling.
+- **A proven trivial fix.** When the reviewer supplied the exact
+  replacement (a JSON tag, a regex, a struct field, a constant) AND you can
+  prove the property directly — a throwaway script, `cat -A`, a marshal
+  check, the one covering test — apply it yourself, run that proof, commit,
+  and ledger `Task <N>: applied <finding> — proof: <what you ran>`. The
+  proof replaces the re-review round; a re-read of a one-line change is
+  weaker evidence than executing it. Anything you cannot prove in one
+  command enters the loop.
+
+Everything else enters the loop. A fix round is one fix dispatch plus one
+scoped re-review. **Three rounds maximum per task** — a real session's
+worst task spent four re-reviews and a fourth round on one task, ~100
+minutes, and converged on adjudication anyway.
+
+**Rounds 1-2 — resume the original implementer.** Send it the open findings
+verbatim. Its context is intact: it knows the task, the code, and its own
+choices. If your harness cannot send another message to a live subagent,
+dispatch a fresh `sdd-implementer` carrying the brief path, the report-file
+path, and the findings — the report file is the persistent memory either way.
+
+**Round 3 — escalate:** dispatch a fresh `sdd-escalation-implementer` with
+the brief path, the report-file path, the open findings, and this framing: "A
+prior implementer attempted this task [N] times; you own it now. Read the
+report file for what was tried." A loop that survives two resumes usually
+means the implementer cannot see its own problem — fresh eyes and a
+capability bump in one move.
+
+**Every round, either way:** the implementer fixes, re-runs the tests
+covering the amended code, appends its fix report to the same report file,
+and returns the short contract. Before re-dispatching the reviewer, confirm
+the fix report contains the covering tests, the command run, and the
+output; dispatch the re-review once all three are present. Name the
+covering test files in the fix message — a one-line fix does not need the
+whole suite.
+
+**The re-review is scoped.** Run `bash scripts/review-package PLAN_FILE FIX_BASE HEAD`
+where FIX_BASE is the head the previous review saw, and dispatch
+`sdd-rereviewer` with [re-review-prompt.md](re-review-prompt.md): the
+findings list, the brief, the report file, and the printed diff path. It
+has no shell and no file writes, and a four-call budget. It yields
+`finding_verdicts[]` (ADDRESSED / NOT_ADDRESSED with file:line),
+`new_breakage[]` for the fix diff only, `out_of_scope[]`, and
+`round_verdict`. New Critical/Important entries in `new_breakage` join the
+open findings list. `out_of_scope` goes to the ledger as deferred minors —
+it never extends the loop.
+
+**After each round,** append to the ledger:
+`Task <N>: fix round <R>/3 (<X> addressed, <Y> open — <finding one-liners>; commits <a7>..<b7>)`
+
+Outside the proven-trivial route above, never fix findings yourself in the
+controller session — your context stays clean for coordination, and an
+unproven controller fix skips review.
+
+**The breaker.** When round 3's re-review still leaves findings open, stop
+dispatching. Adjudicate each open finding yourself — you hold the plan and
+the cross-task context the reviewer lacks:
+
+- **The reviewer is wrong, or the point is contestable:** park it —
+  `Task <N>: parked — <finding> — Ruling: <why the code stands>`. The final
+  review sees both sides.
+- **Real, but nothing downstream builds on it:** park it the same way, with
+  a ruling that says it's real and deferred.
+- **Real and load-bearing** — a later task builds on it, or it reveals a
+  plan defect: rule on the smallest change that unblocks the dependent work,
+  ledger it as `Task <N>: Ruling: <finding> — <what you decided and why>`,
+  and carry it into the next task's dispatch. Parking a structural failure
+  silently lets every dependent task build on it. Stop only when the defect
+  leaves every path forward a guess.
+
+Adjudicate only at the cap. Adjudicating earlier to end a loop is
+pre-judging with a different name. Every adjudication is a ledger entry —
+a silent discard is forbidden.
+
+### 5. Complete the task
+
+When the review comes back clean — or every open finding is parked with a
+ruling at the cap — append the completion line to the ledger in the same
+message as your other bookkeeping:
+
+- `Task <N>: complete (commits <base7>..<head7>, review clean)`
+- `Task <N>: complete (commits <base7>..<head7>, <K> parked)` after a
+  tripped breaker
+
+Then mark the todo complete and move on. Never move to the next task while
+the review has open Critical/Important issues that are neither fixed nor
+parked-with-ruling at the cap.
+
+## Final Review
+
+The final whole-branch review gets a package too: run
+`bash scripts/review-package PLAN_FILE MERGE_BASE HEAD` (MERGE_BASE = the commit the
+branch started from, e.g. `git merge-base main HEAD`) and include the
+printed path in the final review dispatch, so the final reviewer reads
+one file instead of re-deriving the branch diff with git commands. Dispatch
+`sdd-final-reviewer` (xhigh reasoning, once per plan — see Agent
+Selection) with a `task` that names: the package path, the plan/spec
+paths, and the ledger's deferred-minor and parked-with-ruling lines
+verbatim. Do not wrap it in requesting-code-review's `code-reviewer.md` —
+that template asks a prose "Ready to merge?" and has no slot for the
+package or the ledger. The agent file's `<output>` governs: findings with
+file:line, a `### Ledger triage` section (each line → fix-before-merge |
+accept, with reason), a `### Declined to judge` section, and a closing
+**Mergeable:** yes | no. Every "Declined to judge" line is a ruling you
+make and ledger, exactly like a plan conflict — `Final: Ruling: <behavior
+the reviewer set aside> — <what a reasonable person using this software
+gets, and why that stands or why it is now a finding> — <cost if wrong>`;
+one that becomes a finding joins the findings below.
+
+If the final whole-branch review returns findings, dispatch ONE fix subagent
+with the complete findings list — not one fixer per finding.
+Per-finding fixers each rebuild context and re-run suites; a real
+session's final-review fix wave cost more than all its tasks combined.
+Then run exactly one scoped re-review of the fix wave
+(`bash scripts/review-package PLAN_FILE FIX_BASE HEAD` over the fix range,
+[re-review-prompt.md](re-review-prompt.md)).
+Adjudicate any residual findings as in the task loop's breaker: park with
+rulings, or rule on the load-bearing ones and ledger what you decided. Only
+the four classes above stop you here. There is no second fix wave —
+residual load-bearing findings surface to your human partner when
+finishing-a-development-branch presents the options.
+
+## Finish
+
+Before you delete anything, collect every ledger line containing `Ruling:` —
+preflight rulings, parked findings, breaker adjudications, all of them — into
+your final message under "Rulings I made", in the order you made them, each
+with what it costs if wrong. The list is exhaustive: if the ledger holds a
+ruling, the list holds it. That list is the only place the decisions you
+took on your human partner's behalf reach them — they read it and rework
+whatever you got wrong. A ruling that dies with the workspace was a decision
+made in secret.
+
+When the final whole-branch review is clean and its fixes are merged,
+delete this plan's workspace (`rm -rf <workspace>`) — the git history is
+the record now. Sibling directories belong to other plans; leave them
+alone.
+
+Use superpowers:finishing-a-development-branch.
+
+## Common Rationalizations
+
+| Excuse | Reality |
+|--------|---------|
+| "Close enough on spec compliance" | Reviewer found spec gaps = not done. Fix or hit the cap and adjudicate — those are the only exits. |
+| "I'll fix it myself, dispatching is overhead" | Only when the reviewer gave the exact replacement AND you can prove it in one command — then ledger the proof. Anything else: resume the implementer. |
+| "One more round will converge" | Past three rounds, rounds don't converge — the failure is structural. Adjudicate and route. |
+| "The reviewer will just find something new anyway" | Scoped re-reviews verify fixes; they cannot wander. New findings on untouched code go to the ledger, not the loop. |
+| "This finding is obviously wrong, I'll drop it" | You adjudicate only at the cap, and every ruling is a ledger entry. Silent discards are forbidden. |
+| "The fix was small, skip the re-review" | Small is not the test — proven is. If you cannot run one command that proves the exact property (and ledger it), the round ends with a scoped re-review. |
+| "Reviews slow the loop down" | The loop without reviews is just unverified churn. Reviews are the loop's brakes and steering. |
+| "Ledger bookkeeping is overhead" | The ledger is what survives compaction. Controllers without one have re-dispatched entire completed task sequences. |
+| "The implementer spawned its own reviewer — free extra assurance" | It's a duplicate seat reviewing the same diff; the task review is the gate. A worker-spawned reviewer is a defect to flag, not rigor. |
+
+## Example Workflow
+
+```
+You: I'm using Subagent-Driven Development to execute this plan.
+
+[Setup: worktree verified]
+[Read plan file once: docs/superpowers/plans/feature-plan.md]
+[Resolve workspace: bash scripts/sdd-workspace docs/superpowers/plans/feature-plan.md — no ledger inside, fresh start]
+[Create todos for all tasks]
+
+Task 1: Hook installation script
+
+[Run task-brief for Task 1; dispatch implementer with brief + report paths + context]
+
+Implementer: "Before I begin - should the hook be installed at user or system level?"
+
+You: "User level (~/.config/superpowers/hooks/)"
+
+Implementer: [Later]
+  - Implemented install-hook command
+  - Added tests, 5/5 passing
+  - Self-review: Found I missed --force flag, added it
+  - Committed
+
+[Run review-package PLAN_FILE BASE HEAD; dispatch sdd-reviewer with the printed path]
+sdd-reviewer: spec_compliance: compliant · task_quality: approved
+  strengths: ["Good test coverage", "clean"] · findings: []
+
+[Ledger: Task 1: complete (commits a1b2c3d..d4e5f6a, review clean)]
+
+Task 2: Recovery modes
+
+[Run task-brief for Task 2; dispatch implementer with brief + report paths + context]
+
+Implementer: [No questions]
+  - Added verify/repair modes
+  - 8/8 tests passing
+  - Committed
+
+[Run review-package PLAN_FILE BASE HEAD; dispatch sdd-reviewer with the printed path]
+sdd-reviewer: spec_compliance: issues · task_quality: needs_fixes
+  spec_issues: ["Missing: progress reporting (spec says 'report every 100 items')"]
+  findings: [{severity: Important, location: src/recovery.js:7, body: "Magic number (100)"}]
+
+[Fix round 1: resume the implementer with both findings]
+Implementer: Added progress reporting, extracted PROGRESS_INTERVAL constant.
+  Re-ran test/recovery.test.js — 10/10 passing. Fix report appended.
+
+[Run review-package PLAN_FILE FIX_BASE HEAD; dispatch sdd-rereviewer]
+sdd-rereviewer: round_verdict: all_addressed
+  finding_verdicts: [{Missing progress reporting: ADDRESSED, src/recovery.js:41},
+                     {Magic number: ADDRESSED, src/recovery.js:7}]
+  new_breakage: [] · out_of_scope: []
+
+[Ledger: Task 2: fix round 1/3 (2 addressed, 0 open; commits d4e5f6a..b7c8d9e)]
+[Ledger: Task 2: complete (commits d4e5f6a..b7c8d9e, review clean)]
+
+...
+
+[After all tasks]
+[Run review-package PLAN_FILE MERGE_BASE HEAD; dispatch sdd-final-reviewer]
+sdd-final-reviewer: Mergeable: yes. Ledger triage: deferred minors — none block merge.
+
+[Delete this plan's workspace — the record now lives in git]
+
+Done! Using superpowers:finishing-a-development-branch.
+```