feat(lifecycle): revision findings ledger — structured failure feedback, persisted and delivered down the chain (#486)

* feat(lifecycle): revision findings ledger — structured QA/PR/PM/CEO failure feedback, persisted and delivered down the chain

Every bounce used to survive only as flattened prose: rounds overwrote each
other in notes_structured, request_changes persisted nothing, two raw
dev_notes appends were silently destroyed by the next handoff note, and the
dev prompt pointed at fields (qa_notes via evidence(), pm_notes) the API
never delivered. Agents re-interpreted and re-discovered every failure
before they could start fixing it.

- task_review_findings (migration 071, append-only): file/line/severity/
  criterion(AC-id-validated)/expected/actual/fix/evidence per finding, with
  origin (qa|pr_gate|pm|ceo), round, and an open->addressed->verified
  lifecycle (waived reserved); new tasks.pm_notes + PmReviewContent give
  request_changes a structured home
- producers: fail_review/pr_fail/request_changes take findings=[...] (prose
  issues shimmed+merged for one release, deprecation-logged); ceo_reject
  validates its reason (no 500), lands an origin=ceo finding, and bumps
  round+audit on branchless coordination roots; guardrails at the verb
  chokepoint (nudge >5, hard reject >10, field caps, traversal-safe file);
  the dev_notes data-loss appends are removed; new task.request_changes +
  task.ceo_reject audit events close rework attribution
- delivery: qa_notes/pr_reviewer_notes/pm_notes carry the deterministic
  [F-id8] rendering; claim briefings, evidence(), the REVISION_REQUIRED
  spawn prompt, PM triage bounced-blocks, and A2A bodies deliver open
  findings; round-N+1 QA and gate reviewers get the full prior ledger;
  panel Findings tab + bounced-xN chip; metrics pm_rejects/ceo_rejects +
  findings counts; vault task notes render a Findings section (fail-open)
- resolution closes for every origin: i_am_done and submit_up/submit_root
  take resolved_findings gated by FINDINGS_ADDRESSED (owner-gated so a
  stale non-owner PM can never mutate the ledger); pass_review/pr_pass/
  complete verify-stamp same-transaction; ceo_approve stamps best-effort
- 24 real-DB integration tests drive the full loop through the real
  choreographer; full suite 12856 green

* docs: revision findings ledger sweep — CLAUDE.md, map, RAG corpus

- CLAUDE.md: new ledger section + corrected request_changes row
- docs/map/review-findings.md (new subsystem map) + surgical updates to
  task-service/pr-gate-review/metrics-observability/vault/panel maps
- docs/rag: producers' findings contract across qa/pr-reviewer/developer/
  cell-pm/main-pm/ceo role docs (the PM docs were missing request_changes
  entirely), verb references, and a new architecture/review-findings.md
  disambiguating ledger findings from convention findings

* test(e2e): resubmit resolves the pr_fail finding per the ledger contract

The scripted pr_fail revision loop resubmitted submit_up without
resolved_findings — correctly rejected now that FINDINGS_ADDRESSED gates
the PM resubmit verbs (green locally, red only in CI since the e2e suite
skips without ROBOCO_E2E_SMOKE=1). The scripted PM now reads the open
ledger row pr_fail persisted (new open_finding_ids arc helper) and
resolves it on resubmit, asserting the open set drains — exercising the
coordinator half of the new contract end to end.

---------

Co-authored-by: Renn F <rennf93@users.noreply.github.com>

panel/src/components/metrics/delivery-tab.tsx

@@ -204,6 +204,12 @@ function ReworkCard() {
                         <th className="py-1 font-medium text-right">
                           PR fails
                         </th>
+                        <th className="py-1 font-medium text-right">
+                          PM rejects
+                        </th>
+                        <th className="py-1 font-medium text-right">
+                          CEO rejects
+                        </th>
                       </tr>
                     </thead>
                     <tbody>
@@ -216,6 +222,8 @@ function ReworkCard() {
                           <td className="py-1 text-right">{pct(a.rate)}</td>
                           <td className="py-1 text-right">{a.qa_fails}</td>
                           <td className="py-1 text-right">{a.pr_fails}</td>
+                          <td className="py-1 text-right">{a.pm_rejects}</td>
+                          <td className="py-1 text-right">{a.ceo_rejects}</td>
                         </tr>
                       ))}
                     </tbody>
@@ -238,6 +246,12 @@ function ReworkCard() {
                           <ResponsiveTableCardRow label="PR fails">
                             {a.pr_fails}
                           </ResponsiveTableCardRow>
+                          <ResponsiveTableCardRow label="PM rejects">
+                            {a.pm_rejects}
+                          </ResponsiveTableCardRow>
+                          <ResponsiveTableCardRow label="CEO rejects">
+                            {a.ceo_rejects}
+                          </ResponsiveTableCardRow>
                         </div>
                       </ResponsiveTableCard>
                     ))}

panel/src/components/tasks/task-detail/__tests__/tab-findings.test.tsx

@@ -0,0 +1,137 @@
+import { describe, it, expect, vi } from "vitest";
+import { render, screen } from "@testing-library/react";
+import React from "react";
+import { TaskStatus, Team, TaskType, type Task } from "@/types";
+import type { TaskFindingsResponse } from "@/lib/api/tasks";
+
+// Mirrors tab-progress-collapse.test.tsx's shape: mock the data hook
+// directly rather than wiring a real QueryClient — TabFindings only reads
+// the hook's return value.
+const { useTaskFindings } = vi.hoisted(() => ({ useTaskFindings: vi.fn() }));
+
+vi.mock("@/hooks/use-tasks", () => ({ useTaskFindings }));
+
+import { TabFindings } from "../tab-findings";
+
+function buildTask(overrides: Partial<Task> = {}): Task {
+  return {
+    id: "t1",
+    title: "Task",
+    description: "d",
+    status: TaskStatus.NEEDS_REVISION,
+    team: Team.BACKEND,
+    task_type: TaskType.CODE,
+    acceptance_criteria: [],
+    ...overrides,
+  } as unknown as Task;
+}
+
+describe("TabFindings", () => {
+  it("renders the empty state when the task was never bounced", () => {
+    useTaskFindings.mockReturnValue({
+      data: { findings: [], summary: [], total: 0, truncated: false },
+    });
+    render(<TabFindings task={buildTask()} />);
+    expect(
+      screen.getByText("No revision findings recorded yet."),
+    ).toBeInTheDocument();
+  });
+
+  it("groups findings by round and renders severity/status", () => {
+    const response: TaskFindingsResponse = {
+      findings: [
+        {
+          id: "aaaaaaaa-0000-0000-0000-000000000000",
+          task_id: "t1",
+          origin: "qa",
+          round: 2,
+          author_slug: "be-qa",
+          file: "roboco/services/task.py",
+          line: 42,
+          severity: "blocker",
+          criterion: null,
+          expected: "the endpoint returns 404",
+          actual: "the endpoint returns 500",
+          fix: "add a not-found guard",
+          evidence: null,
+          status: "open",
+          addressed_by_commit: null,
+          resolution_note: null,
+          created_at: "2026-07-11T00:00:00Z",
+          updated_at: null,
+        },
+        {
+          id: "bbbbbbbb-0000-0000-0000-000000000000",
+          task_id: "t1",
+          origin: "qa",
+          round: 1,
+          author_slug: "be-qa",
+          file: "roboco/services/task.py",
+          line: 10,
+          severity: "minor",
+          criterion: null,
+          expected: "x",
+          actual: "y",
+          fix: null,
+          evidence: null,
+          status: "verified",
+          addressed_by_commit: "abc1234def",
+          resolution_note: null,
+          created_at: "2026-07-10T00:00:00Z",
+          updated_at: null,
+        },
+      ],
+      summary: [
+        { origin: "qa", open: 1, addressed: 0, verified: 1, waived: 0 },
+      ],
+      total: 2,
+      truncated: false,
+    };
+    useTaskFindings.mockReturnValue({ data: response, isLoading: false });
+    render(<TabFindings task={buildTask()} />);
+
+    expect(screen.getByText("Round 2")).toBeInTheDocument();
+    expect(screen.getByText("Round 1")).toBeInTheDocument();
+    expect(screen.getByText("blocker")).toBeInTheDocument();
+    expect(screen.getByText("minor")).toBeInTheDocument();
+    expect(screen.getByText("abc1234")).toBeInTheDocument();
+    expect(screen.queryByText(/more not shown/)).toBeNull();
+  });
+
+  it("annotates a truncated ledger with the not-shown remainder", () => {
+    const response: TaskFindingsResponse = {
+      findings: [
+        {
+          id: "cccccccc-0000-0000-0000-000000000000",
+          task_id: "t1",
+          origin: "qa",
+          round: 1,
+          author_slug: "be-qa",
+          file: null,
+          line: null,
+          severity: "minor",
+          criterion: null,
+          expected: "x",
+          actual: "y",
+          fix: null,
+          evidence: null,
+          status: "open",
+          addressed_by_commit: null,
+          resolution_note: null,
+          created_at: "2026-07-11T00:00:00Z",
+          updated_at: null,
+        },
+      ],
+      summary: [
+        { origin: "qa", open: 501, addressed: 0, verified: 0, waived: 0 },
+      ],
+      total: 501,
+      truncated: true,
+    };
+    useTaskFindings.mockReturnValue({ data: response, isLoading: false });
+    render(<TabFindings task={buildTask()} />);
+    expect(
+      screen.getByText("… 500 more not shown (501 total)"),
+    ).toBeInTheDocument();
+  });
+});

panel/src/components/tasks/task-detail/index.ts

@@ -10,6 +10,7 @@ export { TabProgress } from "./tab-progress";
 export { TabCommits } from "./tab-commits";
 export { TabNotes } from "./tab-notes";
 export { TabDependencies } from "./tab-dependencies";
+export { TabFindings } from "./tab-findings";
 export { AcceptanceCriteria } from "./acceptance-criteria";
 export { ProgressTimeline } from "./progress-timeline";
 export { CheckpointCard } from "./checkpoint-card";

panel/src/components/tasks/task-detail/tab-findings.tsx

@@ -0,0 +1,169 @@
+"use client";
+
+import { Task } from "@/types";
+import { useTaskFindings } from "@/hooks/use-tasks";
+import type { TaskFinding } from "@/lib/api/tasks";
+import { Card, CardContent } from "@/components/ui/card";
+import { Badge } from "@/components/ui/badge";
+import { Skeleton } from "@/components/ui/skeleton";
+import { ListChecks } from "lucide-react";
+
+interface TabFindingsProps {
+  task: Task;
+}
+
+const SEVERITY_CLASS: Record<string, string> = {
+  blocker: "bg-red-100 text-red-700 dark:bg-red-900 dark:text-red-300",
+  major:
+    "bg-orange-100 text-orange-700 dark:bg-orange-900 dark:text-orange-300",
+  minor:
+    "bg-yellow-100 text-yellow-700 dark:bg-yellow-900 dark:text-yellow-300",
+  nit: "bg-gray-100 text-gray-600 dark:bg-gray-800 dark:text-gray-400",
+};
+
+const STATUS_CLASS: Record<string, string> = {
+  open: "border-red-300 text-red-700 dark:border-red-800 dark:text-red-300",
+  addressed:
+    "border-blue-300 text-blue-700 dark:border-blue-800 dark:text-blue-300",
+  verified:
+    "border-green-300 text-green-700 dark:border-green-800 dark:text-green-300",
+  waived: "border-border text-muted-foreground",
+};
+
+const ORIGIN_LABEL: Record<string, string> = {
+  qa: "QA",
+  pr_gate: "PR Review",
+  pm: "PM",
+  ceo: "CEO",
+};
+
+function FindingCard({ finding }: { finding: TaskFinding }) {
+  return (
+    <Card>
+      <CardContent className="pt-4 space-y-2">
+        <div className="flex flex-wrap items-center gap-2">
+          <Badge className={SEVERITY_CLASS[finding.severity] ?? SEVERITY_CLASS.nit}>
+            {finding.severity}
+          </Badge>
+          <Badge variant="outline" className={STATUS_CLASS[finding.status]}>
+            {finding.status}
+          </Badge>
+          {finding.file && (
+            <code className="text-xs text-muted-foreground">
+              {finding.file}
+              {finding.line != null ? `:${finding.line}` : ""}
+            </code>
+          )}
+          {finding.criterion && (
+            <span className="text-xs text-muted-foreground">
+              · {finding.criterion}
+            </span>
+          )}
+          {finding.addressed_by_commit && (
+            <code className="ml-auto text-xs text-muted-foreground">
+              {finding.addressed_by_commit.slice(0, 7)}
+            </code>
+          )}
+        </div>
+        <div className="space-y-1 text-sm">
+          <p>
+            <span className="text-muted-foreground">Expected:</span>{" "}
+            {finding.expected}
+          </p>
+          <p>
+            <span className="text-muted-foreground">Actual:</span>{" "}
+            {finding.actual}
+          </p>
+          {finding.fix && (
+            <p>
+              <span className="text-muted-foreground">Fix:</span> {finding.fix}
+            </p>
+          )}
+        </div>
+        {finding.evidence && (
+          <details className="text-xs text-muted-foreground">
+            <summary className="cursor-pointer select-none">Evidence</summary>
+            <pre className="mt-1 whitespace-pre-wrap rounded bg-muted/50 p-2">
+              {finding.evidence}
+            </pre>
+          </details>
+        )}
+        {finding.resolution_note && (
+          <p className="text-xs text-muted-foreground">
+            Resolution: {finding.resolution_note}
+          </p>
+        )}
+      </CardContent>
+    </Card>
+  );
+}
+
+export function TabFindings({ task }: TabFindingsProps) {
+  const { data, isLoading } = useTaskFindings(task.id);
+  const findings = data?.findings ?? [];
+
+  if (isLoading) {
+    return (
+      <div className="space-y-3">
+        <Skeleton className="h-24 w-full" />
+        <Skeleton className="h-24 w-full" />
+      </div>
+    );
+  }
+
+  if (findings.length === 0) {
+    return (
+      <div className="py-12 text-center text-muted-foreground">
+        <ListChecks className="mx-auto mb-4 h-12 w-12 opacity-50" />
+        <p>No revision findings recorded yet.</p>
+        <p className="mt-2 text-sm">
+          Findings appear here after the first QA / PR-review / PM / CEO
+          bounce.
+        </p>
+      </div>
+    );
+  }
+
+  // Findings arrive newest-round-first; group while preserving that order.
+  const rounds: { round: number; origin: string; items: TaskFinding[] }[] = [];
+  for (const f of findings) {
+    const last = rounds[rounds.length - 1];
+    if (last && last.round === f.round) last.items.push(f);
+    else rounds.push({ round: f.round, origin: f.origin, items: [f] });
+  }
+
+  return (
+    <div className="space-y-6">
+      {(data?.summary.length ?? 0) > 0 && (
+        <div className="flex flex-wrap gap-2">
+          {data!.summary.map((s) => (
+            <Badge key={s.origin} variant="outline">
+              {ORIGIN_LABEL[s.origin] ?? s.origin}: {s.open} open ·{" "}
+              {s.addressed + s.verified + s.waived} closed
+            </Badge>
+          ))}
+        </div>
+      )}
+      {rounds.map((group) => (
+        <div key={group.round} className="space-y-3">
+          <div className="flex items-center gap-2">
+            <h3 className="text-sm font-semibold">Round {group.round}</h3>
+            <Badge variant="outline">
+              {ORIGIN_LABEL[group.origin] ?? group.origin}
+            </Badge>
+          </div>
+          <div className="space-y-3">
+            {group.items.map((f) => (
+              <FindingCard key={f.id} finding={f} />
+            ))}
+          </div>
+        </div>
+      ))}
+      {data?.truncated && (
+        <p className="text-xs text-muted-foreground">
+          … {data.total - findings.length} more not shown ({data.total} total)
+        </p>
+      )}
+    </div>
+  );
+}

panel/src/components/tasks/task-detail/task-header.tsx

@@ -49,6 +49,7 @@ import {
   Send,
   ThumbsUp,
   ThumbsDown,
+  RotateCcw,
 } from "lucide-react";
 import { toast } from "sonner";
 import { TaskTypeBadge } from "../task-type-badge";
@@ -142,6 +143,7 @@ export function TaskHeader({ task, onAction, nav }: TaskHeaderProps) {
   const overrideStatuses: TaskStatus[] = Object.values(TaskStatus).filter(
     (s) => s !== task.status && !nextStatuses.includes(s),
   );
+  const revisionCount = task.revision_count ?? 0;
   const [deleteOpen, setDeleteOpen] = useState(false);
 
   // Inline editing states
@@ -604,6 +606,22 @@ export function TaskHeader({ task, onAction, nav }: TaskHeaderProps) {
                   </span>
                 </>
               )}
+
+              {/* Bounced chip — hidden for a never-bounced task. */}
+              {revisionCount > 0 && (
+                <Tooltip>
+                  <TooltipTrigger asChild>
+                    <span className="inline-flex h-7 shrink-0 items-center gap-1 rounded-md border border-amber-300 bg-amber-100 px-2 text-xs font-medium text-amber-800 dark:border-amber-800 dark:bg-amber-900 dark:text-amber-300">
+                      <RotateCcw className="h-3 w-3" />
+                      bounced x{revisionCount}
+                    </span>
+                  </TooltipTrigger>
+                  <TooltipContent>
+                    Sent back to needs_revision {revisionCount} time
+                    {revisionCount === 1 ? "" : "s"} — see the Findings tab
+                  </TooltipContent>
+                </Tooltip>
+              )}
             </div>
           </div>
         </div>

panel/src/components/tasks/task-detail/task-tabs.tsx

@@ -2,6 +2,7 @@
 
 import { usePathname, useRouter, useSearchParams } from "next/navigation";
 import { Task } from "@/types";
+import { useTaskFindings } from "@/hooks/use-tasks";
 import { Tabs, TabsContent, TabsList, TabsTrigger } from "@/components/ui/tabs";
 import { Badge } from "@/components/ui/badge";
 import {
@@ -15,6 +16,7 @@ import { TabProgress } from "./tab-progress";
 import { TabCommits } from "./tab-commits";
 import { TabNotes } from "./tab-notes";
 import { TabDependencies } from "./tab-dependencies";
+import { TabFindings } from "./tab-findings";
 import {
   FileText,
   Layout,
@@ -22,6 +24,7 @@ import {
   GitCommit,
   StickyNote,
   Link2,
+  ListChecks,
   type LucideIcon,
 } from "lucide-react";
 
@@ -53,6 +56,8 @@ export function TaskTabs({ task }: TaskTabsProps) {
     (task.auditor_notes ? 1 : 0) +
     (task.quick_context ? 1 : 0);
   const depsCount = task.dependency_ids.length + task.blocker_ids.length;
+  const { data: findingsData } = useTaskFindings(task.id);
+  const findingsCount = findingsData?.total ?? 0;
 
   const tabs: TabDef[] = [
     {
@@ -96,6 +101,13 @@ export function TaskTabs({ task }: TaskTabsProps) {
       hint: "Dependencies and blockers",
       count: depsCount > 0 ? depsCount : undefined,
     },
+    {
+      value: "findings",
+      label: "Findings",
+      icon: ListChecks,
+      hint: "Revision-findings ledger — QA / PR-review / PM / CEO bounce feedback",
+      count: findingsCount > 0 ? findingsCount : undefined,
+    },
   ];
 
   // The active tab lives in the URL (?tab=) so it survives reloads,
@@ -116,7 +128,7 @@ export function TaskTabs({ task }: TaskTabsProps) {
 
   return (
     <Tabs value={activeTab} onValueChange={handleTabChange} className="mt-6">
-      <TabsList className="grid w-full grid-cols-6 lg:w-auto lg:inline-grid">
+      <TabsList className="grid w-full grid-cols-7 lg:w-auto lg:inline-grid">
         {tabs.map((tab) => (
           <Tooltip key={tab.value}>
             <TooltipTrigger asChild>
@@ -154,6 +166,9 @@ export function TaskTabs({ task }: TaskTabsProps) {
         <TabsContent value="dependencies">
           <TabDependencies task={task} />
         </TabsContent>
+        <TabsContent value="findings">
+          <TabFindings task={task} />
+        </TabsContent>
       </div>
     </Tabs>
   );

panel/src/hooks/use-tasks.ts

@@ -28,6 +28,7 @@ export const taskKeys = {
   subtasks: (parentId: string) =>
     [...taskKeys.all, "subtasks", parentId] as const,
   boardReview: (id: string) => [...taskKeys.all, "board-review", id] as const,
+  findings: (id: string) => [...taskKeys.all, "findings", id] as const,
   stats: () => [...taskKeys.all, "stats"] as const,
   statsByTeam: () => [...taskKeys.all, "stats-by-team"] as const,
 };
@@ -68,6 +69,18 @@ export function useBoardReview(taskId: string, enabled = true) {
   });
 }
 
+// The revision-findings ledger for the panel's Findings tab. Cheap enough to
+// fetch eagerly alongside the task (mirrors useBoardReview's shape) so the
+// tab's badge count is available without an extra click.
+export function useTaskFindings(taskId: string) {
+  return useQuery({
+    queryKey: taskKeys.findings(taskId),
+    queryFn: () => tasksApi.getFindings(taskId),
+    enabled: !!taskId,
+    staleTime: 30000,
+  });
+}
+
 export function useSubtasks(parentTaskId: string) {
   return useQuery({
     queryKey: taskKeys.subtasks(parentTaskId),

panel/src/lib/api/tasks.ts

@@ -34,6 +34,47 @@ export interface BoardReviewEntry {
   timestamp: string | null;
 }
 
+// One row of the revision-findings ledger. Matches the backend
+// TaskFindingResponse schema (roboco/api/schemas/tasks.py).
+export interface TaskFinding {
+  id: string;
+  task_id: string;
+  origin: "qa" | "pr_gate" | "pm" | "ceo";
+  round: number;
+  author_slug: string | null;
+  file: string | null;
+  line: number | null;
+  severity: "blocker" | "major" | "minor" | "nit";
+  criterion: string | null;
+  expected: string;
+  actual: string;
+  fix: string | null;
+  evidence: string | null;
+  status: "open" | "addressed" | "verified" | "waived";
+  addressed_by_commit: string | null;
+  resolution_note: string | null;
+  created_at: string;
+  updated_at: string | null;
+}
+
+// Per-origin status counts — matches TaskFindingsSummaryRow.
+export interface TaskFindingsSummaryRow {
+  origin: string;
+  open: number;
+  addressed: number;
+  verified: number;
+  waived: number;
+}
+
+export interface TaskFindingsResponse {
+  findings: TaskFinding[];
+  summary: TaskFindingsSummaryRow[];
+  // Whole-ledger aggregates: the findings list is capped server-side, and
+  // truncated flags when it holds fewer rows than total.
+  total: number;
+  truncated: boolean;
+}
+
 // Wire shape of GET /tasks/summary (backend TaskSummaryResponse) — exactly
 // the fields list views render; everything fat stays on GET /tasks/{id}.
 interface TaskSummaryWire {
@@ -160,6 +201,18 @@ export const tasksApi = {
     return data;
   },
 
+  // The revision-findings ledger (qa_fail/pr_fail/request_changes/ceo_reject),
+  // newest round first, plus per-origin status counts. Empty for a task
+  // never bounced.
+  getFindings: async (taskId: string): Promise<TaskFindingsResponse> => {
+    if (isMockMode())
+      return { findings: [], summary: [], total: 0, truncated: false };
+    const { data } = await api.get<TaskFindingsResponse>(
+      "/tasks/" + taskId + "/findings",
+    );
+    return data;
+  },
+
   // Create task
   create: async (task: TaskCreate): Promise<Task> => {
     if (isMockMode()) {

panel/src/types/index.ts

@@ -276,6 +276,9 @@ export interface Task {
   // Review Status
   self_verified: boolean;
   qa_verified: boolean | null;
+  // Bounces into needs_revision (task-header "bounced xN" chip). Optional:
+  // summaries and mocks omit it.
+  revision_count?: number;
   // Git/Development Context
   branch_name: string | null;
   pr_number: number | null;
@@ -1345,6 +1348,8 @@ export interface AgentReworkRate {
   rate: number;
   qa_fails: number;
   pr_fails: number;
+  pm_rejects: number;
+  ceo_rejects: number;
 }
 
 export interface TeamReworkRate {
@@ -1391,8 +1396,12 @@ export interface TaskMetrics {
   revision_count: number;
   qa_fails: number;
   pr_fails: number;
+  pm_rejects: number;
+  ceo_rejects: number;
   stints: number;
   stages: StageEffort[];
+  findings_open: number;
+  findings_total: number;
 }
 
 export interface MemberScorecard {