fix(recovery): project context evidence for latest-run checks (#17463)

## Thinking Path

> - Paperclip manages AI agents and their tasks.
> - Recovery sweeps decide whether tasks have a valid next step.
> - Those decisions inspect saved run metadata.
> - Latest-run queries also return full prompts and execution setup
data.
> - This repeats large responses when a sweep leaves a task unchanged.
> - This change projects the context evidence recovery actually reads.
> - A row-version check keeps stale evidence from changing a task.

## Linked Issues or Issue Description

**What happened?**

The latest-run recovery queries retrieve large stored contexts for each
candidate.
The decision usually consumes small recovery, message, or workspace-mode
fields.

**Expected behavior**

Keep all decision evidence while selecting a small response. Preserve
complete
stored contexts and full execution reads. Detect concurrent source
changes.

**Steps to reproduce**

1. Store a terminal run with large task Markdown, wake, workspace, and
continuation bodies.
2. Call `reconcileStrandedAssignedIssues` for its assigned task.
3. Observe the full context returned by the latest-run query on each
sweep.

**Paperclip version or commit**

Baseline: `410e5c6eedba98588a87e4104153898fb0437532`.
Related read optimization: Refs #17405. Complementary fix: Refs #17460.
These PRs have no code dependency. Searched existing PRs and issues for
run
projections, recovery reads, and database transfer. No duplicate was
found.

**Deployment mode**

External PostgreSQL. Verification uses disposable real PostgreSQL and
synthetic data.

## What Changed

- Project typed context evidence in all three latest-run recovery
queries.
- Keep only the used workspace mode and external-interaction flag from
nested setup objects.
- Preserve original JSON types, message references, retry evidence, and
existing query scopes and order.
- Read PostgreSQL `xmin` with the run. Require the same row version
during native blocked-wait repair.
- Retain the original full-context comparison for callers that supply a
full run without a row version.
- Test large payloads, malformed JSON shapes, review-message recovery,
and concurrent source edits.

## Verification

- Final-source integration checks: **450 passed** across context
projections, native finalization recovery, and the assembled heartbeat
recovery service.
- Focused issue-lock race check: **9 passed** (401 outside the filter),
including omitted-context and late-output edits and successful repair on
a fresh sweep.
- Final-source recursive typecheck and full build pass. The full local
stable suite is still running; current-head CI must also pass before
merge.
- Real-service SQL capture against the baseline: one returned row falls
from 699,724 to 573 serialized JSON bytes (99.918% smaller).
- The fixture uses four 174,764-byte incompressible execution fields.
All unaffected returned columns compare equal. Regression tests check
the retained decision evidence and unchanged stored records.
- Five alternating warm samples after two warmups: baseline median 1.479
ms, treatment 1.084 ms. Ranges overlap; no request-latency speedup is
claimed. Plans were captured outside the timing interval.
- An initial full-context hash projection cut bytes but raised the local
median to 7.457 ms. The final row-version check removes that cost.
- One integration run overlapped the hash-export rename and failed seven
tests on the removed export. A fresh final-source run passed all 450
tests. The earlier mixed-source failure remains recorded and is not
counted as a pass.
- All current-head CI checks pass on
`17add20e391c30ed0496f4b39eb9c59779efa6b4`. Greptile: **5/5, medium
impact**, no findings or unresolved threads on that head.
- Hosted transfer and product behavior remain to be verified on staging
after merge.

## Risks

The context selection must track its recovery consumers. Tests preserve
typed
flags, retry budgets, external-chat references, and malformed-input
behavior.
The row-version guard is conservative: any intervening run update defers
native
blocker repair to a fresh sweep. Concurrent omitted-context and output
edits are
covered through the real service and issue-lock wait; a fresh sweep must
still repair.
There is no migration, prompt truncation, history deletion, or
authorization cache.

## Model Used

OpenAI GPT-6 through Codex, with source inspection, code execution, and
GitHub tools.
The exact served model ID and context-window size are not exposed in
this session.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

Co-authored-by: Paperclip <noreply@paperclip.ing>

server/src/__tests__/heartbeat-process-recovery.test.ts

@@ -15676,6 +15676,8 @@ describeEmbeddedPostgres("heartbeat orphaned process recovery", () => {
     "comment",
     "wake",
     "run",
+    "run_context",
+    "run_metadata",
     "attention",
   ] as const)(
     "does not clobber a newer %s while native blocker repair waits on the issue lock",
@@ -15750,6 +15752,12 @@ describeEmbeddedPostgres("heartbeat orphaned process recovery", () => {
             continuationPolicy: "wake_assignee",
             payload: { version: 1, questions: [] },
           });
+        if (change === "run_context")
+          await tx.update(heartbeatRuns).set({
+            contextSnapshot: sql`jsonb_set(${heartbeatRuns.contextSnapshot}, '{paperclipTaskMarkdown}', '"Changed execution instructions"'::jsonb)`,
+          }).where(eq(heartbeatRuns.id, runId));
+        if (change === "run_metadata")
+          await tx.update(heartbeatRuns).set({ stdoutExcerpt: "Late execution output" }).where(eq(heartbeatRuns.id, runId));
         [snapshot] = await tx
           .select()
           .from(issues)
@@ -15793,6 +15801,10 @@ describeEmbeddedPostgres("heartbeat orphaned process recovery", () => {
           ),
       ).toHaveLength(0);
       expect(mockAdapterExecute).not.toHaveBeenCalled();
+      if (change === "run_context" || change === "run_metadata") {
+        expect((await heartbeat.reconcileStrandedAssignedIssues()).escalated).toBe(1);
+        expect(mockAdapterExecute).not.toHaveBeenCalled();
+      }
     },
   );
 

server/src/services/recovery/run-context-read.test.ts

@@ -0,0 +1,84 @@
+import { randomBytes, randomUUID } from "node:crypto";
+import { agents, companies, createDb, heartbeatRuns } from "@paperclipai/db";
+import { eq } from "drizzle-orm";
+import { afterAll, beforeAll, describe, expect, it } from "vitest";
+import { getEmbeddedPostgresTestSupport, startEmbeddedPostgresTestDatabase } from "../../__tests__/helpers/embedded-postgres.js";
+import { isExternalChatPresentationContext } from "../heartbeat-run-summary.js";
+import { legacyDispositionEpisode } from "./legacy-continuation.js";
+import { decideIssueReviewPathRecovery } from "./review-path-recovery.js";
+import { recoveryRunContextColumn, recoveryRunVersionColumn } from "./run-context-read.js";
+
+const support = await getEmbeddedPostgresTestSupport();
+
+describe.skipIf(!support.supported)("recovery context PostgreSQL projection", () => {
+  let database: Awaited<ReturnType<typeof startEmbeddedPostgresTestDatabase>>;
+  let db: ReturnType<typeof createDb>;
+  let companyId: string;
+  let agentId: string;
+  beforeAll(async () => {
+    database = await startEmbeddedPostgresTestDatabase("recovery-context-read-");
+    db = createDb(database.connectionString);
+    const [company] = await db.insert(companies).values({ name: "Recovery fixture", issuePrefix: "RC" }).returning();
+    const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Recovery agent", adapterType: "process" }).returning();
+    companyId = company.id;
+    agentId = agent.id;
+  }, 60_000);
+  afterAll(async () => {
+    await db?.$client.end();
+    await database?.cleanup();
+  });
+
+  async function read(contextSnapshot: unknown) {
+    const [run] = await db.insert(heartbeatRuns).values({ companyId, agentId, contextSnapshot: contextSnapshot as Record<string, unknown> | null }).returning();
+    const [selected] = await db.select({ context: recoveryRunContextColumn, version: recoveryRunVersionColumn })
+      .from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
+    return { run, selected };
+  }
+
+  it("retains typed recovery evidence while excluding large execution data", async () => {
+    const issueId = randomUUID();
+    const payload = randomBytes(128 * 1024).toString("base64");
+    const evidence = {
+      issueId, taskId: issueId, retryReason: "issue_continuation_needed", source: "chat:slack",
+      wakeReason: "assignment", handoffReason: "successful_run_missing_state", handoffRequired: true,
+      handoffAttempt: 2, maxHandoffAttempts: 3, sourceRunId: randomUUID(), resumeFromRunId: randomUUID(),
+      missingDisposition: "clear_next_step", providerQuotaRetryNotBefore: "2099-01-01T00:00:00Z",
+      transientRetryNotBefore: "2099-01-02T00:00:00Z", dispositionRepairAttempt: 2,
+      dispositionRepairFingerprint: "fingerprint", legacyDispositionEpisode: { id: "episode", attempt: 1, maxAttempts: 2 },
+      reviewPathRecoveryAttempt: 1, reviewPathConsumedRef: "reference", interactionId: randomUUID(),
+      approvalId: randomUUID(), clearReason: "resolved", monitorAttemptCount: 4,
+      wakeCommentIds: [randomUUID(), randomUUID()], wakeCommentId: randomUUID(), commentId: randomUUID(),
+      externalChatContinuation: true,
+    };
+    const context = { ...evidence, paperclipWake: { externalInteractionContinuation: true, prompt: payload },
+      paperclipWorkspace: { mode: "shared_workspace", instructions: payload }, executionContinuation: { prompt: payload }, paperclipTaskMarkdown: payload };
+    const { run, selected } = await read(context);
+    expect(selected.context).toEqual({ ...evidence, paperclipWake: { externalInteractionContinuation: true }, paperclipWorkspace: { mode: "shared_workspace" } });
+    expect(Buffer.byteLength(JSON.stringify(selected))).toBeLessThan(4096);
+    expect(isExternalChatPresentationContext(selected.context)).toBe(isExternalChatPresentationContext(context));
+    expect(legacyDispositionEpisode({ ...run, contextSnapshot: selected.context })).toEqual(legacyDispositionEpisode(run));
+    const [stored] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
+    expect(stored.contextSnapshot).toEqual(context);
+    const [serialized] = await db.$client.unsafe("SELECT xmin::text AS version FROM heartbeat_runs WHERE id = $1", [run.id]);
+    expect(selected.version).toBe(serialized.version);
+    await db.update(heartbeatRuns).set({ contextSnapshot: { ...context, paperclipTaskMarkdown: "changed" } }).where(eq(heartbeatRuns.id, run.id));
+    const [changed] = await db.select({ context: recoveryRunContextColumn, version: recoveryRunVersionColumn }).from(heartbeatRuns).where(eq(heartbeatRuns.id, run.id));
+    expect(changed.context).toEqual(selected.context);
+    expect(changed.version).not.toBe(selected.version);
+  });
+
+  it.each([null, [], ["unexpected"], "scalar", 42, { issueId: 42, taskId: false, paperclipWake: [], paperclipWorkspace: null }].map(context => ({ context })))("preserves decisions for malformed context $context", async ({ context }) => {
+    const { selected } = await read(context);
+    expect(isExternalChatPresentationContext(selected.context)).toBe(isExternalChatPresentationContext(context));
+    expect(legacyDispositionEpisode({ id: "run", contextSnapshot: selected.context })).toEqual(legacyDispositionEpisode({ id: "run", contextSnapshot: context }));
+  });
+
+  it("retains external-chat review recovery message references and monitor identity", async () => {
+    const context = { source: "chat:slack", wakeReason: "issue_monitor_fired", clearReason: "fired", monitorAttemptCount: 3,
+      wakeCommentIds: [randomUUID(), randomUUID()], prompt: randomBytes(64 * 1024).toString("base64") };
+    const { selected } = await read(context);
+    const input = { issueId: randomUUID(), sourceRunId: randomUUID(), assigneeAgentId: agentId, reviewAttention: { state: "stalled" } as Parameters<typeof decideIssueReviewPathRecovery>[0]["reviewAttention"], existingWake: false };
+    expect(decideIssueReviewPathRecovery({ ...input, contextSnapshot: selected.context }))
+      .toEqual(decideIssueReviewPathRecovery({ ...input, contextSnapshot: context }));
+  });
+});

server/src/services/recovery/run-context-read.ts

@@ -0,0 +1,28 @@
+import { heartbeatRuns } from "@paperclipai/db";
+import { sql } from "drizzle-orm";
+
+const recoveryContextKeys = [
+  "issueId", "taskId", "retryReason", "source", "wakeReason",
+  "handoffReason", "handoffRequired", "handoffAttempt", "maxHandoffAttempts",
+  "sourceRunId", "resumeFromRunId", "missingDisposition",
+  "providerQuotaRetryNotBefore", "transientRetryNotBefore",
+  "dispositionRepairAttempt", "dispositionRepairFingerprint", "legacyDispositionEpisode",
+  "reviewPathRecoveryAttempt", "reviewPathConsumedRef", "interactionId", "approvalId",
+  "clearReason", "monitorAttemptCount", "wakeCommentIds", "wakeCommentId", "commentId",
+  "externalChatContinuation", "paperclipWake", "paperclipWorkspace",
+] as const;
+
+export const recoveryRunContextColumn = sql<Record<string, unknown> | null>`
+  case when ${heartbeatRuns.contextSnapshot} is null then null else (
+    select coalesce(jsonb_object_agg(field.key, case field.key
+      when 'paperclipWake' then jsonb_build_object(
+        'externalInteractionContinuation', field.value->'externalInteractionContinuation')
+      when 'paperclipWorkspace' then jsonb_build_object('mode', field.value->'mode')
+      else field.value end), '{}'::jsonb)
+    from jsonb_each(case when jsonb_typeof(${heartbeatRuns.contextSnapshot}) = 'object'
+      then ${heartbeatRuns.contextSnapshot} else '{}'::jsonb end) as field(key, value)
+    where field.key in (${sql.join(recoveryContextKeys.map(key => sql`${key}`), sql`, `)})
+  ) end
+`;
+
+export const recoveryRunVersionColumn = sql<string>`${heartbeatRuns}.xmin::text`;

server/src/services/recovery/service.ts

@@ -1,3 +1,4 @@
+import { recoveryRunContextColumn, recoveryRunVersionColumn } from "./run-context-read.js";
 import { isAgentAwaitingSetup } from "../../modules/agent-lifecycle/index.js";
 import { onboardingFirstTaskAwaitingUserSql } from "./onboarding-first-task-wait.js";
 import { isAiAuthenticationBlocked } from "../ai-auth-failure.js";
@@ -25,6 +26,7 @@ import {
   desc,
   eq,
   gt,
+  getTableColumns,
   gte,
   inArray,
   isNotNull,
@@ -247,6 +249,7 @@ type LatestIssueRun =
       | "createdAt"
     > & {
       resultJson?: unknown;
+      rowVersion?: string;
     })
   | null;
 type SuccessfulLatestIssueRun = NonNullable<LatestIssueRun> & {
@@ -977,7 +980,8 @@ export function recoveryService(
         status: heartbeatRuns.status,
         error: heartbeatRuns.error,
         errorCode: heartbeatRuns.errorCode,
-        contextSnapshot: heartbeatRuns.contextSnapshot,
+        contextSnapshot: recoveryRunContextColumn,
+        rowVersion: recoveryRunVersionColumn,
         livenessState: heartbeatRuns.livenessState,
         resultJson: heartbeatRuns.resultJson,
         startedAt: heartbeatRuns.startedAt,
@@ -1007,7 +1011,8 @@ export function recoveryService(
         status: heartbeatRuns.status,
         error: heartbeatRuns.error,
         errorCode: heartbeatRuns.errorCode,
-        contextSnapshot: heartbeatRuns.contextSnapshot,
+        contextSnapshot: recoveryRunContextColumn,
+        rowVersion: recoveryRunVersionColumn,
         livenessState: heartbeatRuns.livenessState,
         resultJson: heartbeatRuns.resultJson,
         startedAt: heartbeatRuns.startedAt,
@@ -1618,7 +1623,10 @@ export function recoveryService(
         return false;
 
       const source = await tx
-        .select()
+        .select({
+          ...getTableColumns(heartbeatRuns),
+          rowVersion: recoveryRunVersionColumn,
+        })
         .from(heartbeatRuns)
         .where(
           and(
@@ -1634,8 +1642,9 @@ export function recoveryService(
         source.id !== latestRun.id ||
         source.agentId !== current.assigneeAgentId ||
         source.status !== "succeeded" ||
-        JSON.stringify(source.contextSnapshot) !==
-          JSON.stringify(latestRun.contextSnapshot)
+        (latestRun.rowVersion !== undefined
+          ? source.rowVersion !== latestRun.rowVersion
+          : JSON.stringify(source.contextSnapshot) !== JSON.stringify(latestRun.contextSnapshot))
       )
         return false;
       const proof = await nativeBlockedUnblockAction(
@@ -1903,7 +1912,8 @@ export function recoveryService(
         status: heartbeatRuns.status,
         error: heartbeatRuns.error,
         errorCode: heartbeatRuns.errorCode,
-        contextSnapshot: heartbeatRuns.contextSnapshot,
+        contextSnapshot: recoveryRunContextColumn,
+        rowVersion: recoveryRunVersionColumn,
         livenessState: heartbeatRuns.livenessState,
         resultJson: heartbeatRuns.resultJson,
         startedAt: heartbeatRuns.startedAt,