Repository navigation
[agent-v1 17/18] feat(agent): activation and agent-panel E2E - #16425
CodeJuggernaut wants to merge 14 commits into
Conversation
Lands the slice-17 tail: the migrated agent E2E (the full chat-turn suite with its websocket mocks and the app-mode spec, replacing the shell-era spec and its fixture, whose only importer was that spec) and the README's production-activation documentation at canonical bytes. One cited seam: the mocks' tool-call events adapt to the chain-hardened agent schema (status vocabulary and the required tool_call_id), which supersedes the contract they were written against. The full-union sweep closing the chain found everything else already home: activation wiring, agent error catalog, telemetry events and their tests, and all locale translations landed through the chain's own fix rounds in forms that supersede canonical; the remaining base-canonical deltas are main drift, ruled seams, and receipted fix-forward paths. E2E specs are extracted and lint/typecheck-clean, recorded as not-executed locally (they need a running backend); no CI runs on this branch by design.
Review findings on the slice-17 cut, fixed in place: restore the README's CRDT follower section to its pre-cut bytes (blob 14af8a0) - the cut had regressed it to canonical's ungated-tree text, which misdocuments this tree's followerGate (URL param > persisted opt-in > VITE_AGENT_CRDT_FOLLOWER) and dropped the env from the repro command; restore the flag-off suite's positive anchor and data-agent-gate-settled wait so the negative assertions cannot pass vacuously before the gate runs; port the deleted fixture's t.comfy.org ingest route into the mocks so the fabricated posthog token never emits live third-party traffic; restore the typed empty ListAssetsResponse (total/has_more are schema-required and validateAssetResponse throws on a parse failure); carry forward the dock/undock close-button journey and the shimmer stylesheet-liveness pin from the replaced spec. Declared in the mocks docblock: the suite requires a cloud-distribution build (registerAgentPanelExtension returns early otherwise), and the users/settings page routes do not cover ComfyPage.setupUser's APIRequestContext path. Census corrections: the en-locale delta is not pure key-ordering churn (9 canonical-only keys dropped with no live referents, 14 tip-only, 1 value change - zero agent keys lost across all 14 locales); the sweep's 538-classified denominator is not reproducible from the commit alone (name-status against canonical lists 821 rows; the delta is the ~400-entry customNodes reorg plus per-family grouping). The run-mode surface (store, REST calls, 12 runMode keys) is parked, not lost: removed by ancestor 252468e with the deferral marked in-tree by Composer.test.ts's describe.skip - severity ruling left to the row owner. Dev-mode force-enable makes the flag-off case unmeaningful under test:browser:local (pre-existing); the no-empty-pattern warning at the ws fixture arg is canonical bytes, Playwright-required. Local gates only: no CI runs on this branch by design.
…igation' into hub-17-step # Conflicts: # browser_tests/tests/agent/agentPanelMocks.ts
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Team Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Comment |
🎭 Playwright: ✅ 1888 passed, 0 failed · 1 flaky📊 Browser Reports
📦 Bundle Size
⚡ Performance Report
Absolute values
Raw data{
"timestamp": "2026-09-01T12:40:19.051Z",
"gitSha": "84e077c53f1484c55ce9261f15c2b1f5a388aac9",
"branch": "nathaniel/agent-v1/17-activation-e2e",
"measurements": [
{
"name": "canvas-idle",
"durationMs": 2071.3539999999853,
"styleRecalcs": 9,
"styleRecalcDurationMs": 9.863,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 623.431,
"heapDeltaBytes": 629104,
"heapUsedBytes": 70698704,
"domNodes": -280,
"jsHeapTotalBytes": 4710400,
"scriptDurationMs": 12.607999999999997,
"eventListeners": -185,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-idle",
"durationMs": 2084.7230000000536,
"styleRecalcs": 7,
"styleRecalcDurationMs": 8.19,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 572.3430000000001,
"heapDeltaBytes": 567236,
"heapUsedBytes": 70284304,
"domNodes": -283,
"jsHeapTotalBytes": 5234688,
"scriptDurationMs": 12.100999999999997,
"eventListeners": -185,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "canvas-mouse-sweep",
"durationMs": 1995.552000000032,
"styleRecalcs": 82,
"styleRecalcDurationMs": 51.21300000000001,
"layouts": 14,
"layoutDurationMs": 5.506,
"taskDurationMs": 1006.038,
"heapDeltaBytes": 4715508,
"heapUsedBytes": 74590664,
"domNodes": -280,
"jsHeapTotalBytes": 6807552,
"scriptDurationMs": 164.322,
"eventListeners": -185,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "canvas-mouse-sweep",
"durationMs": 1923.4809999999243,
"styleRecalcs": 79,
"styleRecalcDurationMs": 46.661,
"layouts": 14,
"layoutDurationMs": 5.417,
"taskDurationMs": 935.4529999999999,
"heapDeltaBytes": -7977216,
"heapUsedBytes": 61873188,
"domNodes": -281,
"jsHeapTotalBytes": 5758976,
"scriptDurationMs": 150.731,
"eventListeners": -153,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1678.0449999999973,
"styleRecalcs": 31,
"styleRecalcDurationMs": 19.167000000000005,
"layouts": 6,
"layoutDurationMs": 1.385,
"taskDurationMs": 405.591,
"heapDeltaBytes": 4361400,
"heapUsedBytes": 74209260,
"domNodes": 78,
"jsHeapTotalBytes": 5505024,
"scriptDurationMs": 13.909,
"eventListeners": 19,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1754.9790000000485,
"styleRecalcs": 31,
"styleRecalcDurationMs": 20.703,
"layouts": 6,
"layoutDurationMs": 1.387,
"taskDurationMs": 420.443,
"heapDeltaBytes": 4671092,
"heapUsedBytes": 74297116,
"domNodes": 78,
"jsHeapTotalBytes": 5767168,
"scriptDurationMs": 15.108999999999998,
"eventListeners": 19,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.670000000000012,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "dom-widget-clipping",
"durationMs": 708.2149999999956,
"styleRecalcs": 11,
"styleRecalcDurationMs": 8.786000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 484.57099999999997,
"heapDeltaBytes": 18681920,
"heapUsedBytes": 87787904,
"domNodes": 18,
"jsHeapTotalBytes": 5505024,
"scriptDurationMs": 164.611,
"eventListeners": 2,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "dom-widget-clipping",
"durationMs": 693.8420000000178,
"styleRecalcs": 10,
"styleRecalcDurationMs": 7.259999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 454.537,
"heapDeltaBytes": 18746900,
"heapUsedBytes": 87832188,
"domNodes": 16,
"jsHeapTotalBytes": 5242880,
"scriptDurationMs": 156.08899999999997,
"eventListeners": 2,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-idle",
"durationMs": 2049.3769999999927,
"styleRecalcs": 5,
"styleRecalcDurationMs": 4.818,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 740.6590000000001,
"heapDeltaBytes": -12904324,
"heapUsedBytes": 86879480,
"domNodes": -271,
"jsHeapTotalBytes": -1085440,
"scriptDurationMs": 38.88,
"eventListeners": -155,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-idle",
"durationMs": 2029.81299999999,
"styleRecalcs": 3,
"styleRecalcDurationMs": 2.7840000000000016,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 618.1490000000001,
"heapDeltaBytes": -13581036,
"heapUsedBytes": 87203896,
"domNodes": -271,
"jsHeapTotalBytes": -1871872,
"scriptDurationMs": 33.847,
"eventListeners": -153,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-pan",
"durationMs": 2630.621000000019,
"styleRecalcs": 65,
"styleRecalcDurationMs": 13.571000000000003,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1772.4609999999998,
"heapDeltaBytes": 23256904,
"heapUsedBytes": 88523984,
"domNodes": -275,
"jsHeapTotalBytes": 2060288,
"scriptDurationMs": 820.957,
"eventListeners": -183,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-pan",
"durationMs": 2432.205999999951,
"styleRecalcs": 62,
"styleRecalcDurationMs": 9.769,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1704.1530000000002,
"heapDeltaBytes": 25599892,
"heapUsedBytes": 90908332,
"domNodes": 0,
"jsHeapTotalBytes": 225280,
"scriptDurationMs": 812.225,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "large-graph-zoom",
"durationMs": 3753.4279999999853,
"styleRecalcs": 60,
"styleRecalcDurationMs": 12.042999999999997,
"layouts": 60,
"layoutDurationMs": 15.726999999999999,
"taskDurationMs": 1848.335,
"heapDeltaBytes": 8038196,
"heapUsedBytes": 75681368,
"domNodes": 0,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 825.999,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.670000000000012,
"p95FrameDurationMs": 16.80000000000109
},
{
"name": "large-graph-zoom",
"durationMs": 3607.132999999976,
"styleRecalcs": 60,
"styleRecalcDurationMs": 11.840000000000003,
"layouts": 60,
"layoutDurationMs": 15.232,
"taskDurationMs": 1773.485,
"heapDeltaBytes": 7705772,
"heapUsedBytes": 74972912,
"domNodes": 0,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 778.162,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "legacy-node-drag",
"durationMs": 3211.15199999997,
"styleRecalcs": 40,
"styleRecalcDurationMs": 5.787000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 2106.589,
"heapDeltaBytes": 26827552,
"heapUsedBytes": 93404832,
"domNodes": 0,
"jsHeapTotalBytes": 417792,
"scriptDurationMs": 1044.186,
"eventListeners": 188,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.80000000000109
},
{
"name": "legacy-node-drag",
"durationMs": 3158.525000000054,
"styleRecalcs": 40,
"styleRecalcDurationMs": 6.041999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 2054.017,
"heapDeltaBytes": 21964108,
"heapUsedBytes": 89684356,
"domNodes": 0,
"jsHeapTotalBytes": 417792,
"scriptDurationMs": 997.4130000000001,
"eventListeners": 188,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "minimap-idle",
"durationMs": 2052.9070000000047,
"styleRecalcs": 5,
"styleRecalcDurationMs": 5.847999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 720.8860000000001,
"heapDeltaBytes": 11031152,
"heapUsedBytes": 86921820,
"domNodes": -272,
"jsHeapTotalBytes": -1122304,
"scriptDurationMs": 37.276,
"eventListeners": -185,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "minimap-idle",
"durationMs": 2034.5680000000357,
"styleRecalcs": 5,
"styleRecalcDurationMs": 4.7589999999999995,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 684.6560000000001,
"heapDeltaBytes": -16086204,
"heapUsedBytes": 86491248,
"domNodes": -270,
"jsHeapTotalBytes": -823296,
"scriptDurationMs": 43.545,
"eventListeners": -153,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 620.8359999999971,
"styleRecalcs": 46,
"styleRecalcDurationMs": 10.873000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 429.64399999999995,
"heapDeltaBytes": 14764100,
"heapUsedBytes": 87013460,
"domNodes": 18,
"jsHeapTotalBytes": 5767168,
"scriptDurationMs": 143.85600000000002,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 588.3330000000342,
"styleRecalcs": 47,
"styleRecalcDurationMs": 11.803,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 426.68799999999993,
"heapDeltaBytes": 13642236,
"heapUsedBytes": 85613312,
"domNodes": 20,
"jsHeapTotalBytes": 6291456,
"scriptDurationMs": 143.233,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-idle",
"durationMs": 2018.4319999999616,
"styleRecalcs": 7,
"styleRecalcDurationMs": 7.317,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 594.6550000000001,
"heapDeltaBytes": -6100040,
"heapUsedBytes": 66115072,
"domNodes": -301,
"jsHeapTotalBytes": 5758976,
"scriptDurationMs": 11.086000000000002,
"eventListeners": -201,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-idle",
"durationMs": 2037.527999999952,
"styleRecalcs": 10,
"styleRecalcDurationMs": 10.456999999999997,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 587.2189999999999,
"heapDeltaBytes": 7593584,
"heapUsedBytes": 79952060,
"domNodes": -294,
"jsHeapTotalBytes": 5234688,
"scriptDurationMs": 11.327,
"eventListeners": -231,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1753.1190000000265,
"styleRecalcs": 78,
"styleRecalcDurationMs": 43.151999999999994,
"layouts": 18,
"layoutDurationMs": 6.723,
"taskDurationMs": 913.5930000000001,
"heapDeltaBytes": -10974184,
"heapUsedBytes": 60948232,
"domNodes": -296,
"jsHeapTotalBytes": 5496832,
"scriptDurationMs": 120.93499999999999,
"eventListeners": -201,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1734.6059999999852,
"styleRecalcs": 78,
"styleRecalcDurationMs": 41.134,
"layouts": 18,
"layoutDurationMs": 6.284,
"taskDurationMs": 888.8909999999998,
"heapDeltaBytes": 2018472,
"heapUsedBytes": 74381040,
"domNodes": -296,
"jsHeapTotalBytes": 6283264,
"scriptDurationMs": 113.75099999999999,
"eventListeners": -233,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-transition-enter",
"durationMs": 1400.7589999999936,
"styleRecalcs": 20,
"styleRecalcDurationMs": 34.032999999999994,
"layouts": 16,
"layoutDurationMs": 15.118,
"taskDurationMs": 1027.661,
"heapDeltaBytes": -7344300,
"heapUsedBytes": 87948700,
"domNodes": 13671,
"jsHeapTotalBytes": 10747904,
"scriptDurationMs": 31.28199999999999,
"eventListeners": 2381,
"totalBlockingTimeMs": 141,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "viewport-pan-sweep",
"durationMs": 8541.157,
"styleRecalcs": 246,
"styleRecalcDurationMs": 42.355000000000004,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 6053.664,
"heapDeltaBytes": -19855828,
"heapUsedBytes": 80400988,
"domNodes": -272,
"jsHeapTotalBytes": -1122304,
"scriptDurationMs": 2685.67,
"eventListeners": -137,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "viewport-pan-sweep",
"durationMs": 8698.446999999987,
"styleRecalcs": 247,
"styleRecalcDurationMs": 44.764,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 6147.742,
"heapDeltaBytes": -15814980,
"heapUsedBytes": 84860160,
"domNodes": -269,
"jsHeapTotalBytes": -1908736,
"scriptDurationMs": 2659.456,
"eventListeners": -137,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-idle",
"durationMs": 18477.862000000016,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 18175.586,
"heapDeltaBytes": -25469104,
"heapUsedBytes": 183357000,
"domNodes": -6548,
"jsHeapTotalBytes": 2162688,
"scriptDurationMs": 322.29200000000003,
"eventListeners": -7473,
"totalBlockingTimeMs": 0,
"frameDurationMs": 18.333333333333332,
"p95FrameDurationMs": 16.80000000000291
},
{
"name": "vue-large-graph-idle",
"durationMs": 18492.40299999997,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 18074.846,
"heapDeltaBytes": -28252692,
"heapUsedBytes": 176361940,
"domNodes": -6548,
"jsHeapTotalBytes": 3624960,
"scriptDurationMs": 329.172,
"eventListeners": -7471,
"totalBlockingTimeMs": 0,
"frameDurationMs": 17.776666666666763,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "vue-large-graph-pan",
"durationMs": 23216.178000000014,
"styleRecalcs": 176,
"styleRecalcDurationMs": 27.010999999999953,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 22982.943,
"heapDeltaBytes": -21043624,
"heapUsedBytes": 181051500,
"domNodes": -6548,
"jsHeapTotalBytes": 3063808,
"scriptDurationMs": 935.858,
"eventListeners": -7469,
"totalBlockingTimeMs": 97,
"frameDurationMs": 17.773333333333238,
"p95FrameDurationMs": 16.700000000004366
},
{
"name": "vue-large-graph-pan",
"durationMs": 23341.65500000006,
"styleRecalcs": 180,
"styleRecalcDurationMs": 26.998000000000022,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 23185.076,
"heapDeltaBytes": -12971364,
"heapUsedBytes": 189970424,
"domNodes": -6548,
"jsHeapTotalBytes": 3080192,
"scriptDurationMs": 922.626,
"eventListeners": -7467,
"totalBlockingTimeMs": 196,
"frameDurationMs": 17.77666666666652,
"p95FrameDurationMs": 16.80000000000291
},
{
"name": "vue-node-resize-workload",
"durationMs": 6716.66300000004,
"styleRecalcs": 20,
"styleRecalcDurationMs": 76.22699999999999,
"layouts": 20,
"layoutDurationMs": 27.874999999999996,
"taskDurationMs": 6678.688,
"heapDeltaBytes": -15931460,
"heapUsedBytes": 186976616,
"domNodes": -6548,
"jsHeapTotalBytes": 15208448,
"scriptDurationMs": 132.73600000000002,
"eventListeners": -7471,
"totalBlockingTimeMs": 459,
"frameDurationMs": 18.329999999999927,
"p95FrameDurationMs": 16.80000000000291
},
{
"name": "vue-node-resize-workload",
"durationMs": 6717.857999999978,
"styleRecalcs": 20,
"styleRecalcDurationMs": 79.57300000000001,
"layouts": 20,
"layoutDurationMs": 27.956999999999997,
"taskDurationMs": 6665.01,
"heapDeltaBytes": -11530748,
"heapUsedBytes": 187196812,
"domNodes": -6548,
"jsHeapTotalBytes": 15560704,
"scriptDurationMs": 134.36700000000002,
"eventListeners": -7473,
"totalBlockingTimeMs": 452,
"frameDurationMs": 18.329999999999927,
"p95FrameDurationMs": 16.80000000000291
},
{
"name": "workflow-execution",
"durationMs": 111.35999999999058,
"styleRecalcs": 7,
"styleRecalcDurationMs": 16.485,
"layouts": 2,
"layoutDurationMs": 0.7689999999999999,
"taskDurationMs": 83.86699999999999,
"heapDeltaBytes": 3239220,
"heapUsedBytes": 70266452,
"domNodes": 111,
"jsHeapTotalBytes": 786432,
"scriptDurationMs": 6.682,
"eventListeners": 49,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "workflow-execution",
"durationMs": 464.3019999999751,
"styleRecalcs": 12,
"styleRecalcDurationMs": 22.523999999999997,
"layouts": 3,
"layoutDurationMs": 2.2520000000000007,
"taskDurationMs": 115.296,
"heapDeltaBytes": 5058576,
"heapUsedBytes": 72795624,
"domNodes": 121,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 8.99,
"eventListeners": 99,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.670000000000012,
"p95FrameDurationMs": 16.800000000000182
}
]
}🔬 E2E Coverage
Top 10 files by uncovered lines
|
Codecov Report✅ All modified and coverable lines are covered by tests. @@ Coverage Diff @@
## nathaniel/agent-v1/16-canvas-navigation #16425 +/- ##
========================================================================
Coverage 83.26% 83.27%
========================================================================
Files 2025 2025
Lines 117495 117495
Branches 33209 33210 +1
========================================================================
+ Hits 97838 97839 +1
+ Misses 19304 19303 -1
Partials 353 353
Flags with carried forward coverage won't be shown. Click here to find out more. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
🔍 Cursor Review — Consolidated panel
Triggered by @christian-byrne.
Found 10 finding(s).
| Severity | Count |
|---|---|
| 🟡 Medium | 3 |
| 🟢 Low | 5 |
| ⚪ Nit | 2 |
Panel: 8/8 reviewers contributed findings.
| // (two CI runs held opposite stable states at the same point), so the | ||
| // journey ensures visibility instead of pinning either lifecycle. | ||
| async function ensureOpen(summary: Locator): Promise<void> { | ||
| if ((await summary.getAttribute('aria-expanded')) !== 'true') { |
There was a problem hiding this comment.
🟡 Medium — ensureOpen reads aria-expanded once and then clicks based on that now-stale read, with no retry: if the group auto-expands between the read and the click (the WebSocket events pushed around each call re-render the group), the click collapses it and the following toHaveAttribute('aria-expanded', 'true') can never recover — it just burns the expect timeout. Poll click-and-verify as a unit so it is retry-safe, and since the comment concedes the settled expand state is runtime-order dependent (two CI runs held opposite stable states), track that product nondeterminism rather than only normalizing it away in the test.
Raised by 4 of 8 reviewers (claude-opus-5-thinking-max adversarial, claude-opus-5-thinking-max edge-case, kimi-k3-max adversarial, gemini-3.1-pro edge-case).
| total: 0, | ||
| has_more: false | ||
| } | ||
| await page.route('**/api/assets**', (r) => r.fulfill(jsonRoute(emptyAssets))) |
There was a problem hiding this comment.
🟡 Medium — mockAgentBoot routes assets, billing, features and telemetry but not GET /api/agent/threads or GET /api/workflows, which mounting the panel triggers via refreshHistory() and refreshCloudWorkflowIds(). Those requests escape to whatever backend is running: a 404 goes through surfaceAgentError into the global error overlay that no test dismisses, and a backend that implements them leaks real thread data into the suite. renameFocus.spec.ts routes **/agent/threads for exactly this load path — add both routes here to keep the suite hermetic.
Raised by 1 of 8 reviewers (kimi-k3-max edge-case).
| * early on non-cloud builds, so against a default build every flag-on test | ||
| * fails and the flag-off test passes vacuously. | ||
| * | ||
| * The `/api/users` and `/api/settings` routes serve only the page's own |
There was a problem hiding this comment.
🟡 Medium — The new docstring frames the /api/settings route as harmless because ComfyPage.setupSettings() writes through the APIRequestContext, but that is precisely the problem: the page's own boot GET is intercepted and answered with the mock's two keys, so everything the shared fixture seeds is silently discarded — including Comfy.OnboardingCoachmarks.Seen, which exists so an auto-opened tour overlay cannot swallow clicks. Now that these tests run through comfyPage they inherit that gap; merge the fixture-seeded settings into the mocked response instead of replacing them.
Raised by 1 of 8 reviewers (claude-opus-5-thinking-max edge-case).
|
|
||
| const ws = await getWebSocket() | ||
| await sendButton.click() | ||
| await expect.poll(() => postedMessages.length).toBeGreaterThanOrEqual(1) |
There was a problem hiding this comment.
🟢 Low — toBeGreaterThanOrEqual(1) makes a duplicate-send regression undetectable: a Send click that POSTs the turn twice satisfies this and every later assertion, since only postedMessages[0] is inspected. The sibling test at line 370 polls toBe(1), so the exact bound is available and works — use it here too unless there is a known reason the count can legitimately exceed one.
Raised by 1 of 8 reviewers (claude-opus-5-thinking-max adversarial).
| await panel.getByRole('button', { name: enMessages.agent.stop }).click() | ||
| await expect( | ||
| panel.getByRole('button', { name: enMessages.g.edit }) | ||
| ).toHaveCount(0) |
There was a problem hiding this comment.
🟢 Low — toHaveCount(0) succeeds on its first poll, so this assertion resolves immediately after the Stop click and cannot distinguish "Edit correctly stays hidden until the turn is finalized" from "the cancel has not been processed yet" — as written it can never fail. Anchor it on an observable post-Stop state first (the Stop button reverting to Send, or a cancelling indicator), then assert Edit is still absent.
Raised by 2 of 8 reviewers (claude-opus-5-thinking-max adversarial, claude-opus-5-thinking-max edge-case).
| 'inserting a prompt must not POST a message' | ||
| ).toHaveLength(0) | ||
|
|
||
| const ws = await getWebSocket() |
There was a problem hiding this comment.
🟢 Low — The WebSocketRoute handle is captured once and reused for all ten pushEvent calls in this 30s test, but the fixture only tracks the most recently routed /ws socket. If the app reconnects mid-turn, ws.send() is fire-and-forget against a dead route, so every later event is silently dropped and the failure surfaces as an unrelated assertion timeout far from the cause. Re-resolving await getWebSocket() at each push (as line 380 already does) keeps it pinned to the live socket.
Raised by 2 of 8 reviewers (claude-opus-5-thinking-max adversarial, claude-opus-5-thinking-max edge-case).
| expect( | ||
| postedMessages, | ||
| 'inserting a prompt must not POST a message' | ||
| ).toHaveLength(0) |
There was a problem hiding this comment.
🟢 Low — This is a synchronous read of postedMessages, but a POST triggered by the chip click arrives asynchronously through the route handler on a different path than the preceding toHaveValue round-trip, so the "inserting a prompt must not POST a message" guard can pass while the regression it protects against is present. await expect.poll(() => postedMessages.length).toBe(0) would actually hold it. The same check at line 40 in the flag-off test runs before any interaction at all and is unconditionally true.
Raised by 2 of 8 reviewers (claude-opus-5-thinking-max edge-case, kimi-k3-max edge-case).
| await expect(longestItem).toBeVisible() | ||
| await expect | ||
| .poll(() => menu.boundingBox().then((box) => box?.width)) | ||
| .toBeGreaterThan(186) |
There was a problem hiding this comment.
🟢 Low — This test pins sub-pixel layout geometry: a hardcoded 186px width floor plus a <= 1px symmetry comparison of fractional boundingBox() reads, both driven by font metrics that differ between CI (Linux) and developer machines and shift with any font or locale change. It also assumes longestItem.locator('span').first() is the icon and that addFromAssets is the longest label, neither of which is enforced, so a label change turns this into a silent false pass or a failure that gets triaged as a flake.
Raised by 1 of 8 reviewers (claude-opus-5-thinking-max adversarial).
| tool_name: 'set_widget', | ||
| status: 'success', | ||
| tool_call_id: 'tool-call-1', | ||
| args: ['workflow', 'set-widget', 'workflow.json'], |
There was a problem hiding this comment.
⚪ Nit — args is not declared by zAgentToolCallData — it only type-checks because that schema is .passthrough() — and nothing under src/workbench/extensions/agent reads it, so this positional array's shape is validated by nothing and can drift from what the backend actually sends. Either model args in the schema or drop it until a consumer exists, so the mock keeps the typed discipline the rest of the file follows.
Raised by 2 of 8 reviewers (claude-opus-5-thinking-max adversarial, claude-opus-5-thinking-max edge-case).
| // The bootstrapped project token makes PostHogTelemetryProvider run a real | ||
| // posthog.init(); route its ingest host so CI never emits live third-party | ||
| // traffic under the fabricated token. | ||
| await page.route('**://t.comfy.org/**', (r) => |
There was a problem hiding this comment.
⚪ Nit — The comment claims CI never emits live third-party traffic, but this route enumerates one host rather than denying by default; the guarantee holds only because agentFeatures() happens to omit the other provider tokens and posthog_api_host, so PostHog falls back to t.comfy.org. A catch-all abort for non-localhost origins would make the stated invariant enforced instead of incidental.
Raised by 1 of 8 reviewers (claude-opus-5-thinking-max adversarial).
The edit button sits in an action row that stays pointer-events-none until the message group is hovered, so the click retries against the wrapper until timeout. Hover first, like a user would.
There was a problem hiding this comment.
🔍 Cursor Review — Consolidated panel
Triggered by @christian-byrne.
Round 2 — ledger: 10 prior finding(s) across 1 round(s) (10 never answered).
Found 10 finding(s).
| Severity | Count |
|---|---|
| 🟡 Medium | 3 |
| 🟢 Low | 6 |
| ⚪ Nit | 1 |
Panel: 8/8 reviewers contributed findings.
| total: 0, | ||
| has_more: false | ||
| } | ||
| await page.route('**/api/assets**', (r) => r.fulfill(jsonRoute(emptyAssets))) |
There was a problem hiding this comment.
🟡 Medium — mockAgentBoot routes assets, billing, features and telemetry but still not GET /api/agent/threads or GET /api/workflows, both of which mounting the panel triggers via refreshHistory() and refreshCloudWorkflowIds(). Those requests escape to whatever backend the run points at: a 404 goes through surfaceAgentError into a global error overlay that no test dismisses, and a backend that implements them pulls real thread data into the suite. renameFocus.spec.ts already routes **/agent/threads for exactly this load path.
Raised by 4 of 8 reviewers (claude-opus-5-thinking-max adversarial, gemini-3.1-pro adversarial, gemini-3.1-pro edge-case, kimi-k3-max edge-case).
| * early on non-cloud builds, so against a default build every flag-on test | ||
| * fails and the flag-off test passes vacuously. | ||
| * | ||
| * The `/api/users` and `/api/settings` routes serve only the page's own |
There was a problem hiding this comment.
🟡 Medium — The new docstring only rules out the write direction: ComfyPage.setupSettings does go through the APIRequestContext, but the page's own boot GET of /api/settings is intercepted here and answered with the mock's two keys, so everything the shared comfyPage fixture seeds is silently discarded — including Comfy.OnboardingCoachmarks.Seen, which exists so an auto-opened tour overlay cannot swallow clicks. Merge the fixture-seeded settings into the mocked response instead of replacing them.
Raised by 3 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case, kimi-k3-max edge-case).
| // (two CI runs held opposite stable states at the same point), so the | ||
| // journey ensures visibility instead of pinning either lifecycle. | ||
| async function ensureOpen(summary: Locator): Promise<void> { | ||
| if ((await summary.getAttribute('aria-expanded')) !== 'true') { |
There was a problem hiding this comment.
🟡 Medium — ensureOpen reads aria-expanded once and then clicks based on that now-stale read with no retry. If the group auto-expands between the read and the click — and the comment directly above concedes the settled expand state is runtime-order dependent — the click collapses it and the following toHaveAttribute('aria-expanded', 'true') can never recover; it just burns the expect timeout. Poll click-and-verify as a unit so it is retry-safe.
Raised by 3 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case, kimi-k3-max edge-case).
| 'inserting a prompt must not POST a message' | ||
| ).toHaveLength(0) | ||
|
|
||
| const ws = await getWebSocket() |
There was a problem hiding this comment.
🟢 Low — The WebSocketRoute handle is captured once here and reused for all ten pushEvent calls in this 30s test, and getWebSocket() itself caches the first routed socket without invalidating it on close. If the app reconnects mid-turn, ws.send() is fire-and-forget against a dead route, so every later event is silently dropped and the failure surfaces as an unrelated assertion timeout far from the cause. Re-resolving await getWebSocket() at each push (as line 380 already does) keeps it pinned to the live socket.
Raised by 3 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case, claude-opus-5-thinking-max edge-case).
| await panel.getByRole('button', { name: enMessages.agent.stop }).click() | ||
| await expect( | ||
| panel.getByRole('button', { name: enMessages.g.edit }) | ||
| ).toHaveCount(0) |
There was a problem hiding this comment.
🟢 Low — toHaveCount(0) succeeds on its first poll, so this assertion resolves immediately after the Stop click and cannot distinguish "Edit correctly stays hidden until the turn is finalized" from "the cancel has not been processed yet" — as written it can never fail. Anchor it on an observable post-Stop state first (Stop reverting to Send, or a cancelling indicator), then assert Edit is still absent.
Raised by 2 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case).
| scrollbarColor: getComputedStyle(element).scrollbarColor | ||
| })) | ||
|
|
||
| expect(track.backgroundColor).toBe('rgba(0, 0, 0, 0)') |
There was a problem hiding this comment.
🟢 Low — getComputedStyle(element, '::-webkit-scrollbar-track') falls back to the element's own computed style when no rule matches that pseudo-element, so backgroundColor === 'rgba(0, 0, 0, 0)' and backgroundImage === 'none' hold for any plain transparent div. These two assertions therefore cannot distinguish "the transparent-track rule is applied" from "the rule does not exist at all", which is the regression they are meant to catch; only the element-level scrollbarColor read on line 311 carries real signal.
Raised by 1 of 8 reviewers (claude-opus-5-thinking-max edge-case).
|
|
||
| const ws = await getWebSocket() | ||
| await sendButton.click() | ||
| await expect.poll(() => postedMessages.length).toBeGreaterThanOrEqual(1) |
There was a problem hiding this comment.
🟢 Low — toBeGreaterThanOrEqual(1) makes a duplicate-send regression undetectable: a Send click that POSTs the turn twice satisfies this and every later assertion, since only postedMessages[0] is inspected. The sibling test at line 370 polls toBe(1), so the exact bound is available and works — use it here too unless there is a known reason the count can legitimately exceed one.
Raised by 2 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case).
| expect( | ||
| postedMessages, | ||
| 'inserting a prompt must not POST a message' | ||
| ).toHaveLength(0) |
There was a problem hiding this comment.
🟢 Low — This is a synchronous read of postedMessages, but a POST triggered by the chip click arrives asynchronously through the route handler on a different path than the preceding toHaveValue round-trip, so the "inserting a prompt must not POST a message" guard can pass while the regression it protects against is present. await expect.poll(() => postedMessages.length).toBe(0) would actually hold it. The same check at line 40 in the flag-off test runs before any interaction at all and is unconditionally true.
Raised by 2 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case).
| await expect(longestItem).toBeVisible() | ||
| await expect | ||
| .poll(() => menu.boundingBox().then((box) => box?.width)) | ||
| .toBeGreaterThan(186) |
There was a problem hiding this comment.
🟢 Low — This test pins sub-pixel layout geometry: a hardcoded 186px width floor plus a <= 1px symmetry comparison of fractional boundingBox() reads, both driven by font metrics that differ between CI (Linux) and developer machines and shift with any font or locale change. It also assumes longestItem.locator('span').first() is the icon and that addFromAssets is the longest label, neither of which is enforced, so a label change turns this into a silent false pass or a failure that gets triaged as a flake.
Raised by 2 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case).
| tool_name: 'set_widget', | ||
| status: 'success', | ||
| tool_call_id: 'tool-call-1', | ||
| args: ['workflow', 'set-widget', 'workflow.json'], |
There was a problem hiding this comment.
⚪ Nit — args is not declared by zAgentToolCallData — it only type-checks because that schema is .passthrough() — and nothing under src/workbench/extensions/agent reads it, so this positional array's shape is validated by nothing and can drift from what the backend actually sends. Either model args in the schema or drop it until a consumer exists, so the mock keeps the typed discipline the rest of the file follows.
Raised by 2 of 8 reviewers (gemini-3.1-pro adversarial, gemini-3.1-pro edge-case).
|
This slice is reference-only. Accepted work is routed to #16340, #16477, #16495, and #16498. Full context for agent readersPer Christian's main-target close ruling, #16425 (17 — activation and E2E) is no longer a landing vehicle. It remains readable implementation and review provenance. Activation, browser, follower-gate, and graph-load responsibilities are separated. If a genuinely unaccounted delta is found, reopen this source or file a bounded main-based carrier rather than reviving the stacked merge path. |
Slice 17 of the agent-v1 extraction ladder (stacked on #16424) - the chain tail.
Activation and E2E coverage: the agent panel browser suite (agentPanel.spec, agentPanelAppMode.spec) with hardened tool-call mock frames, plus the rename-focus fence riding from slice 15. First CI execution of the migrated agent E2E happens on this PR.
Already restacked onto current main - the diff shows only this slice's own delta.
Review-audit deltas (salvage matrix): items found during the slicing and review audits that did not make it through the integration merge (#16198).
Activation/E2E cut: the agent panel browser suites with hardened wire mocks - first CI execution of the migrated agent E2E ran on this PR.