Skip to content

Open-source AI browser agent for Chrome and Firefox (monorepo) 🧠

fix: share cold UI snapshot reads across concurrent callers

Emre Sokullu37125e85a8e4
diff --git a/CHANGELOG.md b/CHANGELOG.md
index a7bb519da..d00dc9f09 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -11,7 +11,7 @@ This changelog was generated from the repository Git history and release tags. V
- General DOM monitoring repairs prevent decorative icons, passive updates and acknowledged agent preparation from incorrectly invalidating unchanged controls, while retaining checks for real target or user changes.
- Mid-run task changes, Stop, durable chat recovery, Telegram/WhatsApp observation, recipient ownership and completion receipts are covered by causal and installed-browser regressions.
- File download/upload ownership, literal editor input, exact calculation and complete literal table delivery are repaired across Chrome and Firefox.
-- The final unfiltered suite passed 9,765 Node/main cases and all 21 installed browser stages. Coverage: 65.16% lines/statements, 78.74% branches, 87.25% functions. Both browser builds passed.
+- The original audit passed 9,765 Node/main cases and all 21 installed browser stages; coverage was 65.16% lines/statements, 78.74% branches and 87.25% functions. The Codex review follow-up additionally passed 72 focused, 155 neighbouring and 2,450 main regression cases. Both browser builds passed.
- Real-account two-hour endurance remains unverified. The first fresh PR CodeQL findings check flagged 32 changed-code alerts (31 high, one medium); the comparison and directly related QA fixture repair are recorded in the appendix. This entry records the proposed release; the audited runtime remains 40.0.8.
The complete dated audit, including earlier failures, tracked effects and verification limits, is copied in the [40.2.0 appendix at the end of this changelog](#appendix-4020-quality-audit). A standalone copy is available in [the audit report](docs/quality-audit-2026-10-07.md).
@@ -23,6 +23,7 @@ The complete dated audit, including earlier failures, tracked effects and verifi
- Added a full-suite Node coverage command and optional live-model browser checks.
### Fixed
+- Share concurrent cold UI snapshot reads so service-worker restart and continuation cannot discard saved events or progress; preserve clear/new-run invalidation.
- Reject injected URL parameters in the disposable workflow QA server without changing valid fixtures or weakening replay checks.
- Keep Enter-dispatch uncertainty truthful when reusing an owned chat draft, without weakening pending-send or duplicate-send guards.
- Preserve exact native commit acknowledgement provenance for independent completion reconciliation after interrupted observation.
@@ -4943,3 +4944,29 @@ One newly flagged condition exposed a concrete injection path in the disposable
The whole installed workflow test completed once with exit 0 at 2026-10-11 02:28:34 UTC. **All five cases passed**, including six once-navigated malicious/unknown query browser/server checks and the original four workflow replay, Stop and uncertainty cases. All 34 original assertion-containing lines are retained. All 12 valid preset/default HTML combinations remain byte-identical. The 1,182 product files, fixture, package, lock and index stayed stable during execution; provider fallback requests and external calls were zero. The fixture SHA-256 is `02766f9c98ea758197586c49e213ca5ec85bb54c4c4b1901e714da3f41eab87c`. Local proof: `ci/artifacts/quality-audit-2026-10-07/publication-workflow-fixture-injection-fix-4d8e4c59-82d3-4f15-b0fc-27dbf4eac751/fix-proof-final.json`, SHA-256 `1cb662d4da3ccd38428a1398bb44dbfcb4ab6e73ba71861d031b0c976895fe43`.
This follow-up changes the workflow fixture and audit documentation only. The previously completed unfiltered C8 run and its coverage percentages remain dated evidence for the frozen audited product; they are not relabelled as a new full-suite run after the fixture edit. The initial PR-head security failure remains recorded. Further remote checks may run on the follow-up commit, and their actual results must be reported separately. Real-account two-hour endurance and the bug-free objective remain unqualified.
+
+#### PR #408 Codex review follow-up — October 11
+
+Codex's first review, on `50c8d73a`, identified concurrent cold UI snapshot reads cancelling one another. An `agent_run_state` read overlapping `beginContinuationRunUiSnapshot` after a service-worker restart could make the continuation treat a saved journal as absent and discard its progress and events. The correction is general and applies to both Chrome and Firefox.
+
+`getRunUiSnapshot` now shares one in-flight storage read per tab. Starting a new journal or clearing a conversation still invalidates that read; a late result respects a newer live journal and cannot resurrect cleared state. A failed shared read releases its slot so a later read can retry. Only the paired snapshot-reader functions changed in shipping code.
+
+Meaningful regression proof: with the added tests and original code, **64/72 passed and eight failed**; after the fix, **72/72 passed**, with no skips or cancellations. The cold continuation reproduction previously replaced persisted sequence 2 with sequence 0. It now preserves sequence 2, acknowledged sequence 1 and the retained event, then appends sequence 3. The tests also cover both original read completion orders, shared-read failure/retry and the existing clear/new-run fences.
+
+Validation on the corrected source:
+
+- `node --test test/run-ui-scheduled-lifecycle.mjs`: **72 passed**.
+- Ten neighbouring persistence, startup, continuation, transport, CDP restart/clear and assistant UI test files: **155 passed**, with no failures, skips or cancellations.
+- `node test/run.js`: **2,450 passed, zero failed**.
+- Normal `npm run build:all`: both browser builds passed; all 1,182 source/build files matched.
+- Syntax and whitespace checks passed. Main remained `6da190593be3990dceb83396fb42b9e779ef9f1e`, with no new upstream commits or merge conflicts. The 29 unrelated local files were preserved.
+
+The full 9,765-case coverage run above belongs to the original audited source, before this follow-up. It was not rerun for this focused change. These results do not add a real-account endurance or fresh model-reliability claim. Runtime version remains 40.0.8.
+
+Local evidence: `ci/artifacts/quality-audit-2026-10-07/pr408-cold-run-ui-snapshot-7f40172c-9f97-48e2-bdf7-6d22bf10069f/green-result.json` (SHA-256 `5f1f32a8e5a650dda11171f861dc8787981be3db90b01ff0e36b7a587e1a437f`) and the neighbouring/main logs under `ci/artifacts/pr408-review-cycle/`. Private artifacts are not published.
+
+The completed scan of the preceding exact head `fa49ffef` now confirms workflow fixture alert **#234 fixed**. Its PR inventory contains 59 open alerts and one fixed alert among the same 60 IDs; no new alert IDs appeared. JavaScript analysis returned 59 findings without analysis errors, while Python and Actions returned zero. The separate findings check still reports **31 changed-code annotations: 30 high and one medium**.
+
+The bounded review finds no additional confirmed actionable regression in the 18 test/report/builder annotations: current paths use text-node assignment, escaped or numeric output, HTML-safe JSON serialization, trusted fixture parsing or intentional operator-authored CLI regexes. The 13 shipping annotations retain their inherited conditional trust/fidelity assessments; their flagged statements are also present on main, but that comparison does not close vulnerabilities. No findings were dismissed or waived, and this is not a security-clean claim.
+
+This correction requires a fresh Codex review. Merge readiness must use that review and the resulting head's checks; the earlier review and scan do not qualify a later revision.
diff --git a/docs/quality-audit-2026-10-07.md b/docs/quality-audit-2026-10-07.md
index 78c9e35a2..dde7c6615 100644
--- a/docs/quality-audit-2026-10-07.md
+++ b/docs/quality-audit-2026-10-07.md
@@ -331,3 +331,29 @@ One newly flagged condition exposed a concrete injection path in the disposable
The whole installed workflow test completed once with exit 0 at 2026-10-11 02:28:34 UTC. **All five cases passed**, including six once-navigated malicious/unknown query browser/server checks and the original four workflow replay, Stop and uncertainty cases. All 34 original assertion-containing lines are retained. All 12 valid preset/default HTML combinations remain byte-identical. The 1,182 product files, fixture, package, lock and index stayed stable during execution; provider fallback requests and external calls were zero. The fixture SHA-256 is `02766f9c98ea758197586c49e213ca5ec85bb54c4c4b1901e714da3f41eab87c`. Local proof: `ci/artifacts/quality-audit-2026-10-07/publication-workflow-fixture-injection-fix-4d8e4c59-82d3-4f15-b0fc-27dbf4eac751/fix-proof-final.json`, SHA-256 `1cb662d4da3ccd38428a1398bb44dbfcb4ab6e73ba71861d031b0c976895fe43`.
This follow-up changes the workflow fixture and audit documentation only. The previously completed unfiltered C8 run and its coverage percentages remain dated evidence for the frozen audited product; they are not relabelled as a new full-suite run after the fixture edit. The initial PR-head security failure remains recorded. Further remote checks may run on the follow-up commit, and their actual results must be reported separately. Real-account two-hour endurance and the bug-free objective remain unqualified.
+
+## PR #408 Codex review follow-up — October 11
+
+Codex's first review, on `50c8d73a`, identified concurrent cold UI snapshot reads cancelling one another. An `agent_run_state` read overlapping `beginContinuationRunUiSnapshot` after a service-worker restart could make the continuation treat a saved journal as absent and discard its progress and events. The correction is general and applies to both Chrome and Firefox.
+
+`getRunUiSnapshot` now shares one in-flight storage read per tab. Starting a new journal or clearing a conversation still invalidates that read; a late result respects a newer live journal and cannot resurrect cleared state. A failed shared read releases its slot so a later read can retry. Only the paired snapshot-reader functions changed in shipping code.
+
+Meaningful regression proof: with the added tests and original code, **64/72 passed and eight failed**; after the fix, **72/72 passed**, with no skips or cancellations. The cold continuation reproduction previously replaced persisted sequence 2 with sequence 0. It now preserves sequence 2, acknowledged sequence 1 and the retained event, then appends sequence 3. The tests also cover both original read completion orders, shared-read failure/retry and the existing clear/new-run fences.
+
+Validation on the corrected source:
+
+- `node --test test/run-ui-scheduled-lifecycle.mjs`: **72 passed**.
+- Ten neighbouring persistence, startup, continuation, transport, CDP restart/clear and assistant UI test files: **155 passed**, with no failures, skips or cancellations.
+- `node test/run.js`: **2,450 passed, zero failed**.
+- Normal `npm run build:all`: both browser builds passed; all 1,182 source/build files matched.
+- Syntax and whitespace checks passed. Main remained `6da190593be3990dceb83396fb42b9e779ef9f1e`, with no new upstream commits or merge conflicts. The 29 unrelated local files were preserved.
+
+The full 9,765-case coverage run above belongs to the original audited source, before this follow-up. It was not rerun for this focused change. These results do not add a real-account endurance or fresh model-reliability claim. Runtime version remains 40.0.8.
+
+Local evidence: `ci/artifacts/quality-audit-2026-10-07/pr408-cold-run-ui-snapshot-7f40172c-9f97-48e2-bdf7-6d22bf10069f/green-result.json` (SHA-256 `5f1f32a8e5a650dda11171f861dc8787981be3db90b01ff0e36b7a587e1a437f`) and the neighbouring/main logs under `ci/artifacts/pr408-review-cycle/`. Private artifacts are not published.
+
+The completed scan of the preceding exact head `fa49ffef` now confirms workflow fixture alert **#234 fixed**. Its PR inventory contains 59 open alerts and one fixed alert among the same 60 IDs; no new alert IDs appeared. JavaScript analysis returned 59 findings without analysis errors, while Python and Actions returned zero. The separate findings check still reports **31 changed-code annotations: 30 high and one medium**.
+
+The bounded review finds no additional confirmed actionable regression in the 18 test/report/builder annotations: current paths use text-node assignment, escaped or numeric output, HTML-safe JSON serialization, trusted fixture parsing or intentional operator-authored CLI regexes. The 13 shipping annotations retain their inherited conditional trust/fidelity assessments; their flagged statements are also present on main, but that comparison does not close vulnerabilities. No findings were dismissed or waived, and this is not a security-clean claim.
+
+This correction requires a fresh Codex review. Merge readiness must use that review and the resulting head's checks; the earlier review and scan do not qualify a later revision.
diff --git a/src/chrome/src/background.js b/src/chrome/src/background.js
index 68905bb9b..c9892022c 100644
--- a/src/chrome/src/background.js
+++ b/src/chrome/src/background.js
@@ -2025,26 +2025,32 @@ function askCompletionSucceededForBadge(result, updates = [], error = null) {
async function getRunUiSnapshot(tabId) {
const live = runUiJournal.get(tabId);
if (live) return live;
+ const pendingRead = runUiSnapshotReads.get(tabId);
+ if (pendingRead) return pendingRead.promise;
const read = {};
+ // Cold readers share ownership; only begin/clear may invalidate this read.
+ read.promise = Promise.resolve().then(async () => {
+ try {
+ const pending = runUiPersistenceQueues.get(tabId);
+ if (pending) await pending;
+ const current = runUiJournal.get(tabId);
+ if (current) return current;
+ if (runUiSnapshotReads.get(tabId) !== read) return null;
+ const key = RUN_UI_PREFIX + tabId;
+ const stored = await chrome.storage.session?.get(key);
+ const latest = runUiJournal.get(tabId);
+ if (latest) return latest;
+ if (runUiSnapshotReads.get(tabId) !== read) return null;
+ const snapshot = stored?.[key];
+ if (snapshot && typeof snapshot === 'object') {
+ return runUiJournal.restore(tabId, snapshot);
+ }
+ } catch {}
+ finally { if (runUiSnapshotReads.get(tabId) === read) runUiSnapshotReads.delete(tabId); }
+ return null;
+ });
runUiSnapshotReads.set(tabId, read);
- try {
- const pending = runUiPersistenceQueues.get(tabId);
- if (pending) await pending;
- const current = runUiJournal.get(tabId);
- if (current) return current;
- if (runUiSnapshotReads.get(tabId) !== read) return null;
- const key = RUN_UI_PREFIX + tabId;
- const stored = await chrome.storage.session?.get(key);
- const latest = runUiJournal.get(tabId);
- if (latest) return latest;
- if (runUiSnapshotReads.get(tabId) !== read) return null;
- const snapshot = stored?.[key];
- if (snapshot && typeof snapshot === 'object') {
- return runUiJournal.restore(tabId, snapshot);
- }
- } catch {}
- finally { if (runUiSnapshotReads.get(tabId) === read) runUiSnapshotReads.delete(tabId); }
- return null;
+ return read.promise;
}
function clearRunUiSnapshot(tabId) {
diff --git a/src/firefox/src/background.js b/src/firefox/src/background.js
index 45d1cc78f..77b6ffea0 100644
--- a/src/firefox/src/background.js
+++ b/src/firefox/src/background.js
@@ -2222,26 +2222,32 @@ function askCompletionSucceededForBadge(result, updates = [], error = null) {
async function getRunUiSnapshot(tabId) {
const live = runUiJournal.get(tabId);
if (live) return live;
+ const pendingRead = runUiSnapshotReads.get(tabId);
+ if (pendingRead) return pendingRead.promise;
const read = {};
+ // Cold readers share ownership; only begin/clear may invalidate this read.
+ read.promise = Promise.resolve().then(async () => {
+ try {
+ const pending = runUiPersistenceQueues.get(tabId);
+ if (pending) await pending;
+ const current = runUiJournal.get(tabId);
+ if (current) return current;
+ if (runUiSnapshotReads.get(tabId) !== read) return null;
+ const key = RUN_UI_PREFIX + tabId;
+ const stored = await browser.storage.session?.get(key);
+ const latest = runUiJournal.get(tabId);
+ if (latest) return latest;
+ if (runUiSnapshotReads.get(tabId) !== read) return null;
+ const snapshot = stored?.[key];
+ if (snapshot && typeof snapshot === 'object') {
+ return runUiJournal.restore(tabId, snapshot);
+ }
+ } catch {}
+ finally { if (runUiSnapshotReads.get(tabId) === read) runUiSnapshotReads.delete(tabId); }
+ return null;
+ });
runUiSnapshotReads.set(tabId, read);
- try {
- const pending = runUiPersistenceQueues.get(tabId);
- if (pending) await pending;
- const current = runUiJournal.get(tabId);
- if (current) return current;
- if (runUiSnapshotReads.get(tabId) !== read) return null;
- const key = RUN_UI_PREFIX + tabId;
- const stored = await browser.storage.session?.get(key);
- const latest = runUiJournal.get(tabId);
- if (latest) return latest;
- if (runUiSnapshotReads.get(tabId) !== read) return null;
- const snapshot = stored?.[key];
- if (snapshot && typeof snapshot === 'object') {
- return runUiJournal.restore(tabId, snapshot);
- }
- } catch {}
- finally { if (runUiSnapshotReads.get(tabId) === read) runUiSnapshotReads.delete(tabId); }
- return null;
+ return read.promise;
}
function clearRunUiSnapshot(tabId) {
diff --git a/test/run-ui-scheduled-lifecycle.mjs b/test/run-ui-scheduled-lifecycle.mjs
index 208bca382..a8a23bd36 100644
--- a/test/run-ui-scheduled-lifecycle.mjs
+++ b/test/run-ui-scheduled-lifecycle.mjs
@@ -48,7 +48,7 @@ function backgroundHarness({ Journal, source, runUiUpdatesForCurrentTask }, opti
});
context.browser = context.chrome;
const functions = [
- section(source, 'function beginRunUiSnapshot(', 'async function beginContinuationRunUiSnapshot('),
+ section(source, 'function beginRunUiSnapshot(', 'function recordRunUiEvent('),
section(source, 'function recordRunUiEvent(', 'function isClarificationRequiredRunUpdate('),
section(source, 'function isClarificationRequiredRunUpdate(', 'function finishRunUiSnapshot('),
section(source, 'function finishRunUiSnapshot(', '// Mirror the sidepanel'),
@@ -684,6 +684,87 @@ for (const item of builds) {
}
});
+ for (const reverse of [false, true]) {
+ test(`${item.build}: concurrent cold snapshot readers preserve the same stored journal (${reverse ? 'last' : 'first'} read resolves first)`, async () => {
+ const h = backgroundHarness(item);
+ h.journal.begin(77, 'cold-owned-request', { mode: 'act' });
+ h.journal.record(77, 'cold-owned-request', 'run_status', { status: 'thinking' });
+ h.journal.record(77, 'cold-owned-request', 'text_delta', { text: 'Preserved owned output.' });
+ const saved = structuredClone(h.stored.get(77));
+ h.journal.clear(77);
+ const reads = [];
+ h.context.chrome.storage.session.get = () => new Promise(resolve => { reads.push(resolve); });
+ const first = h.context.getRunUiSnapshot(77);
+ const second = h.context.getRunUiSnapshot(77);
+ await flushMicrotasks();
+ const storageReadCount = reads.length;
+ for (const release of reverse ? [...reads].reverse() : reads) {
+ release({ 'runUi:77': structuredClone(saved) });
+ await flushMicrotasks();
+ }
+ const results = await Promise.all([first, second]);
+ assert.ok(results.every(result => result?.requestId === 'cold-owned-request'));
+ assert.equal(results[0], results[1]);
+ assert.equal(h.journal.get(77), results[0]);
+ assert.equal(results[0].seq, saved.seq);
+ assert.deepEqual(results[0].events, saved.events);
+ assert.equal(storageReadCount, 1, 'readers of the same cold journal share its storage read');
+ assert.equal(h.context.runUiSnapshotReads.size, 0);
+ });
+ }
+
+ test(`${item.build}: a remount read cannot discard persisted events during a cold same-request continuation`, async () => {
+ const h = backgroundHarness(item);
+ h.journal.begin(77, 'cold-continuation-request', { mode: 'act', foreground: true });
+ h.journal.record(77, 'cold-continuation-request', 'run_status', { status: 'thinking' });
+ h.journal.record(77, 'cold-continuation-request', 'plan_review', { planId: 'retained-owned-plan' });
+ h.journal.acknowledge(77, 'cold-continuation-request', 1);
+ const saved = structuredClone(h.stored.get(77));
+ h.journal.clear(77);
+ const reads = [];
+ h.context.chrome.storage.session.get = () => new Promise(resolve => { reads.push(resolve); });
+ const continuation = h.context.beginContinuationRunUiSnapshot(77, 'cold-continuation-request', { kind: 'continue' });
+ await flushMicrotasks();
+ const remount = h.context.getRunUiSnapshot(77);
+ await flushMicrotasks();
+ for (const release of reads) {
+ release({ 'runUi:77': structuredClone(saved) });
+ await flushMicrotasks();
+ }
+ const [resumed, observed] = await Promise.all([continuation, remount]);
+ assert.equal(resumed.seq, 2, 'continuation preserves the persisted sequence instead of beginning a new journal');
+ assert.equal(resumed.ackedSeq, 1);
+ assert.equal(resumed.foreground, true);
+ assert.deepEqual(resumed.events, saved.events);
+ assert.equal(observed, resumed);
+ const next = h.context.recordRunUiEvent(77, 'cold-continuation-request', 'text_delta', { text: 'Continued owned output.' });
+ assert.equal(next.seq, 3);
+ assert.deepEqual(resumed.events.map(event => event.seq), [2, 3]);
+ assert.equal(reads.length, 1);
+ });
+
+ test(`${item.build}: a failed shared cold read permits a later storage retry`, async () => {
+ const h = backgroundHarness(item);
+ const rejects = [];
+ h.context.chrome.storage.session.get = () => new Promise((_resolve, reject) => { rejects.push(reject); });
+ const first = h.context.getRunUiSnapshot(77);
+ const second = h.context.getRunUiSnapshot(77);
+ await flushMicrotasks();
+ for (const reject of rejects) reject(new Error('Owned storage read failed.'));
+ assert.deepEqual(await Promise.all([first, second]), [null, null]);
+ assert.equal(h.context.runUiSnapshotReads.size, 0);
+ let retries = 0;
+ h.context.chrome.storage.session.get = async key => {
+ retries++;
+ return { [key]: { requestId: 'retry-owned-request', status: 'running', events: [], seq: 4 } };
+ };
+ const recovered = await h.context.getRunUiSnapshot(77);
+ assert.equal(recovered.requestId, 'retry-owned-request');
+ assert.equal(recovered.seq, 4);
+ assert.equal(retries, 1);
+ assert.equal(rejects.length, 1);
+ });
+
test(`${item.build}: clearing a conversation during stored UI read cannot resurrect its previous success`, async () => {
const h = backgroundHarness(item);
let release;