Repository navigation
feat(agent): instrument starter-prompt clicks, and attribute the send they produce - #18991
Conversation
🎨 Storybook: ✅ Built — View Storybook🎭 Playwright: ✅ 2160 passed, 0 failed📊 Browser Reports
🌾 Fallow: ✅ PassedFallow audits the diff for duplication, complexity and dead code. It gates on new findings only, so when the audit completes, a red is something this PR introduced or moved — never inherited backlog. A red can also mean the audit itself did not finish; the details below say which happened. Fallow gates on new findings in changed files only, so everything below was introduced or moved by this PR (27 changed files).
Pre-existing findings are baselined in Reproduce locally# the base the gate compared against
npx fallow audit --root . --quiet --format json \
--changed-since 25c268595d612e2ee0aa1749191eee6a022c601f
Full docs: ADR-DEVEX-LINT-0015 📦 Bundle: 10 MB gzip 🔴 +346 kBDetailsSummary
Category Glance App Entry Points — 3.77 kB (baseline 3.73 kB) • 🔴 +38 BMain entry bundles and manifests
Status: 1 added / 1 removed Graph Workspace — 1.46 MB (baseline 1.43 MB) • 🔴 +29.7 kBGraph editor runtime, canvas, workflow orchestration
Status: 2 added / 2 removed / 1 unchanged Views & Navigation — 133 kB (baseline 124 kB) • 🔴 +9.54 kBTop-level views, pages, and routed surfaces
Status: 17 added / 16 removed / 1 unchanged Panels & Settings — 643 kB (baseline 621 kB) • 🔴 +21.7 kBConfiguration panels, inspectors, and settings screens
Status: 28 added / 27 removed User & Accounts — 39.4 kB (baseline 27.8 kB) • 🔴 +11.6 kBAuthentication, profile, and account management bundles
Status: 9 added / 9 removed / 2 unchanged Editors & Dialogs — 126 kB (baseline 126 kB) • 🔴 +691 BModals, dialogs, drawers, and in-app editors
Status: 8 added / 8 removed UI Components — 125 kB (baseline 117 kB) • 🔴 +8.07 kBReusable component library chunks
Status: 16 added / 15 removed Data & Services — 4.06 MB (baseline 3.6 MB) • 🔴 +463 kBStores, services, APIs, and repositories
Status: 15 added / 16 removed Utilities & Hooks — 582 kB (baseline 559 kB) • 🔴 +23.2 kBHelpers, composables, and utility bundles
Status: 31 added / 30 removed / 6 unchanged Vendor & Third-Party — 17.4 MB (baseline 18.1 MB) • 🟢 -684 kBExternal libraries and shared vendor chunks
Status: 12 added / 12 removed / 6 unchanged Other — 17.9 MB (baseline 16.4 MB) • 🔴 +1.54 MBBundles that do not match a named category
Status: 178 added / 174 removed / 122 unchanged ⚡ Performance Report
Show regressions
All metrics
Historical variance (last 15 runs)
Trend (last 15 commits on main)
Raw data{
"timestamp": "2026-09-29T21:06:16.005Z",
"gitSha": "e0d84f346b491173db257fd9518b38481c258c73",
"branch": "feat/agent-starter-prompt-instrumentation",
"measurements": [
{
"name": "canvas-idle",
"durationMs": 2033.4300000000098,
"styleRecalcs": 12,
"styleRecalcDurationMs": 5.240999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 481.471,
"heapDeltaBytes": -2281572,
"heapUsedBytes": 76447768,
"domNodes": -228,
"jsHeapTotalBytes": 3932160,
"scriptDurationMs": 10.682,
"eventListeners": -164,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.6655737704918,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-idle",
"durationMs": 2009.123000000045,
"styleRecalcs": 11,
"styleRecalcDurationMs": 3.8390000000000004,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 451.54999999999995,
"heapDeltaBytes": -21217900,
"heapUsedBytes": 77573176,
"domNodes": -228,
"jsHeapTotalBytes": 3932160,
"scriptDurationMs": 8.551000000000004,
"eventListeners": -164,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666115702479335,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-mouse-sweep",
"durationMs": 1826.9789999999944,
"styleRecalcs": 77,
"styleRecalcDurationMs": 17.927,
"layouts": 12,
"layoutDurationMs": 4.724,
"taskDurationMs": 844.534,
"heapDeltaBytes": -33686588,
"heapUsedBytes": 64766444,
"domNodes": -228,
"jsHeapTotalBytes": 4194304,
"scriptDurationMs": 170.745,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665137614678898,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-mouse-sweep",
"durationMs": 1826.255999999944,
"styleRecalcs": 79,
"styleRecalcDurationMs": 18.049999999999997,
"layouts": 12,
"layoutDurationMs": 4.226999999999999,
"taskDurationMs": 841.6629999999999,
"heapDeltaBytes": -34070968,
"heapUsedBytes": 64033392,
"domNodes": -228,
"jsHeapTotalBytes": 4980736,
"scriptDurationMs": 171.737,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666055045871556,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1739.8770000000354,
"styleRecalcs": 32,
"styleRecalcDurationMs": 10.28,
"layouts": 6,
"layoutDurationMs": 1.191,
"taskDurationMs": 328.30400000000003,
"heapDeltaBytes": 1442548,
"heapUsedBytes": 80468408,
"domNodes": 53,
"jsHeapTotalBytes": 3670016,
"scriptDurationMs": 11.309999999999997,
"eventListeners": 15,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666346153846156,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1731.9820000000163,
"styleRecalcs": 32,
"styleRecalcDurationMs": 11.158000000000001,
"layouts": 6,
"layoutDurationMs": 1.131,
"taskDurationMs": 353.39500000000004,
"heapDeltaBytes": 1470852,
"heapUsedBytes": 80355680,
"domNodes": 53,
"jsHeapTotalBytes": 3670016,
"scriptDurationMs": 13.828999999999997,
"eventListeners": 15,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666346153846156,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "dom-widget-clipping",
"durationMs": 690.9659999999462,
"styleRecalcs": 13,
"styleRecalcDurationMs": 4.043000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 454.169,
"heapDeltaBytes": 18317964,
"heapUsedBytes": 97386608,
"domNodes": 0,
"jsHeapTotalBytes": 3932160,
"scriptDurationMs": 173.49900000000002,
"eventListeners": 0,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665853658536566,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "dom-widget-clipping",
"durationMs": 684.8499999999831,
"styleRecalcs": 12,
"styleRecalcDurationMs": 4.484999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 455.377,
"heapDeltaBytes": -1798200,
"heapUsedBytes": 96986540,
"domNodes": 0,
"jsHeapTotalBytes": 7864320,
"scriptDurationMs": 171.163,
"eventListeners": 0,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66500000000001,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "dom-widget-order-dense-n200-w200-i120",
"durationMs": 1161.5700000000402,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 844.4609999999999,
"heapDeltaBytes": 10367872,
"heapUsedBytes": 85699524,
"domNodes": 0,
"jsHeapTotalBytes": 5668864,
"scriptDurationMs": 73.246,
"eventListeners": 0,
"totalBlockingTimeMs": 543,
"frameDurationMs": 35.35151515151514,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "dom-widget-order-dense-n200-w200-i120",
"durationMs": 1191.4359999999533,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 843.307,
"heapDeltaBytes": 6507168,
"heapUsedBytes": 83635952,
"domNodes": 0,
"jsHeapTotalBytes": 3833856,
"scriptDurationMs": 70.98199999999999,
"eventListeners": 0,
"totalBlockingTimeMs": 548,
"frameDurationMs": 34.8029411764706,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "dom-widget-order-sparse-n1000-w20-i120",
"durationMs": 863.0239999999958,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 484.46500000000015,
"heapDeltaBytes": 16390104,
"heapUsedBytes": 89878940,
"domNodes": 0,
"jsHeapTotalBytes": 266240,
"scriptDurationMs": 174.27300000000002,
"eventListeners": 0,
"totalBlockingTimeMs": 232,
"frameDurationMs": 25.485294117647058,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "dom-widget-order-sparse-n1000-w20-i120",
"durationMs": 879.4800000000578,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 492.50099999999986,
"heapDeltaBytes": 14312816,
"heapUsedBytes": 89421704,
"domNodes": 0,
"jsHeapTotalBytes": 266240,
"scriptDurationMs": 177.91600000000003,
"eventListeners": 0,
"totalBlockingTimeMs": 239,
"frameDurationMs": 25.237142857142835,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "extension-clean-combined-draw",
"durationMs": 10066.233000000011,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 9118.500000000002,
"heapDeltaBytes": 10221304,
"heapUsedBytes": 91570440,
"domNodes": 0,
"jsHeapTotalBytes": -12582912,
"scriptDurationMs": 7740.265,
"eventListeners": 0,
"totalBlockingTimeMs": 1415,
"frameDurationMs": 35.4443661971831,
"p95FrameDurationMs": 66.70000000000073
},
{
"name": "extension-clean-combined-draw",
"durationMs": 10786.323999999922,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 9761.536,
"heapDeltaBytes": 2269120,
"heapUsedBytes": 82931092,
"domNodes": 0,
"jsHeapTotalBytes": -14155776,
"scriptDurationMs": 8036.799,
"eventListeners": 0,
"totalBlockingTimeMs": 1184,
"frameDurationMs": 31.621407624633424,
"p95FrameDurationMs": 66.60000000000036
},
{
"name": "extension-rgthree-prototype-draw",
"durationMs": 10791.434000000037,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 9810.005000000001,
"heapDeltaBytes": 11047600,
"heapUsedBytes": 92594904,
"domNodes": 0,
"jsHeapTotalBytes": -12320768,
"scriptDurationMs": 8146.876000000001,
"eventListeners": 0,
"totalBlockingTimeMs": 1197,
"frameDurationMs": 32.38108108108108,
"p95FrameDurationMs": 66.69999999999891
},
{
"name": "extension-rgthree-prototype-draw",
"durationMs": 11874.164000000064,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 10856.151,
"heapDeltaBytes": 3335508,
"heapUsedBytes": 83981612,
"domNodes": 0,
"jsHeapTotalBytes": -2883584,
"scriptDurationMs": 9102.047,
"eventListeners": 0,
"totalBlockingTimeMs": 1874,
"frameDurationMs": 34.39478260869566,
"p95FrameDurationMs": 66.70000000000073
},
{
"name": "large-graph-idle",
"durationMs": 2054.817000000014,
"styleRecalcs": 7,
"styleRecalcDurationMs": 2.620000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 540.081,
"heapDeltaBytes": -36218400,
"heapUsedBytes": 83152068,
"domNodes": -228,
"jsHeapTotalBytes": -884736,
"scriptDurationMs": 32.94899999999999,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665853658536584,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-idle",
"durationMs": 2042.7359999999908,
"styleRecalcs": 8,
"styleRecalcDurationMs": 3.2279999999999984,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 590.106,
"heapDeltaBytes": -35942364,
"heapUsedBytes": 83374084,
"domNodes": -228,
"jsHeapTotalBytes": -1409024,
"scriptDurationMs": 34.294000000000004,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666393442622944,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-pan",
"durationMs": 2438.3129999999937,
"styleRecalcs": 67,
"styleRecalcDurationMs": 8.821,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1663.602,
"heapDeltaBytes": 27532320,
"heapUsedBytes": 89945024,
"domNodes": -228,
"jsHeapTotalBytes": 225280,
"scriptDurationMs": 880.487,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.78,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-pan",
"durationMs": 2687.852000000021,
"styleRecalcs": 70,
"styleRecalcDurationMs": 9.673000000000002,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1716.68,
"heapDeltaBytes": 21265256,
"heapUsedBytes": 83554172,
"domNodes": -228,
"jsHeapTotalBytes": 225280,
"scriptDurationMs": 902.7310000000001,
"eventListeners": -164,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665838509316774,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "large-graph-zoom",
"durationMs": 3653.9080000000013,
"styleRecalcs": 63,
"styleRecalcDurationMs": 10.790999999999999,
"layouts": 60,
"layoutDurationMs": 13.354,
"taskDurationMs": 1710.411,
"heapDeltaBytes": 9245000,
"heapUsedBytes": 73170820,
"domNodes": 0,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 830.796,
"eventListeners": 0,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666363636363638,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-zoom",
"durationMs": 3679.0079999999534,
"styleRecalcs": 64,
"styleRecalcDurationMs": 10.647,
"layouts": 60,
"layoutDurationMs": 13.455,
"taskDurationMs": 1710.6050000000002,
"heapDeltaBytes": 9476604,
"heapUsedBytes": 73262004,
"domNodes": 0,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 796.473,
"eventListeners": 0,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666063348416294,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "legacy-node-drag",
"durationMs": 3174.5879999999715,
"styleRecalcs": 42,
"styleRecalcDurationMs": 5.547,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1974.988,
"heapDeltaBytes": 26666944,
"heapUsedBytes": 90830880,
"domNodes": 0,
"jsHeapTotalBytes": 417792,
"scriptDurationMs": 1038.4799999999998,
"eventListeners": 180,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.753968253968253,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "legacy-node-drag",
"durationMs": 3113.211999999976,
"styleRecalcs": 43,
"styleRecalcDurationMs": 5.354999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1920.1460000000002,
"heapDeltaBytes": 24191776,
"heapUsedBytes": 88692032,
"domNodes": 0,
"jsHeapTotalBytes": 417792,
"scriptDurationMs": 978.2689999999999,
"eventListeners": 180,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.755675675675683,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "minimap-idle",
"durationMs": 2043.8429999999812,
"styleRecalcs": 6,
"styleRecalcDurationMs": 2.326999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 541.0359999999998,
"heapDeltaBytes": -37311060,
"heapUsedBytes": 83764128,
"domNodes": -228,
"jsHeapTotalBytes": -884736,
"scriptDurationMs": 31.124,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66639344262295,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "minimap-idle",
"durationMs": 2028.8829999999507,
"styleRecalcs": 8,
"styleRecalcDurationMs": 3.2900000000000014,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 468.3539999999999,
"heapDeltaBytes": -37977120,
"heapUsedBytes": 83414784,
"domNodes": -228,
"jsHeapTotalBytes": -98304,
"scriptDurationMs": 35.833000000000006,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666393442622944,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 617.8140000000099,
"styleRecalcs": 48,
"styleRecalcDurationMs": 6.820000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 404.45400000000006,
"heapDeltaBytes": 13338292,
"heapUsedBytes": 93816544,
"domNodes": 0,
"jsHeapTotalBytes": 4456448,
"scriptDurationMs": 148.53900000000002,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.664864864864875,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 618.682999999919,
"styleRecalcs": 46,
"styleRecalcDurationMs": 5.425999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 393.637,
"heapDeltaBytes": 13314164,
"heapUsedBytes": 93241836,
"domNodes": 0,
"jsHeapTotalBytes": 3670016,
"scriptDurationMs": 145.759,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66486486486485,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-idle",
"durationMs": 2040.3830000000198,
"styleRecalcs": 10,
"styleRecalcDurationMs": 4.296999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 440.997,
"heapDeltaBytes": -18488360,
"heapUsedBytes": 62114488,
"domNodes": -244,
"jsHeapTotalBytes": -786432,
"scriptDurationMs": 8.128,
"eventListeners": -188,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66557377049181,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-idle",
"durationMs": 2040.0569999999334,
"styleRecalcs": 9,
"styleRecalcDurationMs": 3.6520000000000006,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 449.561,
"heapDeltaBytes": -9787792,
"heapUsedBytes": 71515856,
"domNodes": -244,
"jsHeapTotalBytes": -524288,
"scriptDurationMs": 7.884000000000002,
"eventListeners": -188,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.80245901639344,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1694.2589999999882,
"styleRecalcs": 73,
"styleRecalcDurationMs": 20.923,
"layouts": 16,
"layoutDurationMs": 5.954,
"taskDurationMs": 670.39,
"heapDeltaBytes": -22794328,
"heapUsedBytes": 58208908,
"domNodes": -244,
"jsHeapTotalBytes": 262144,
"scriptDurationMs": 112.08100000000002,
"eventListeners": -188,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1745.492000000013,
"styleRecalcs": 73,
"styleRecalcDurationMs": 18.360000000000003,
"layouts": 16,
"layoutDurationMs": 5.981999999999999,
"taskDurationMs": 677.658,
"heapDeltaBytes": -2925120,
"heapUsedBytes": 77927348,
"domNodes": -244,
"jsHeapTotalBytes": 262144,
"scriptDurationMs": 111.22999999999999,
"eventListeners": -210,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665714285714284,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-transition-enter",
"durationMs": 1385.4849999999033,
"styleRecalcs": 18,
"styleRecalcDurationMs": 28.919,
"layouts": 5,
"layoutDurationMs": 12.856999999999998,
"taskDurationMs": 920.6780000000001,
"heapDeltaBytes": 21467472,
"heapUsedBytes": 93944924,
"domNodes": 13667,
"jsHeapTotalBytes": 3670016,
"scriptDurationMs": 42.678,
"eventListeners": 2383,
"totalBlockingTimeMs": 132,
"frameDurationMs": 19.17808219178082,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "viewport-pan-sweep",
"durationMs": 8445.79699999997,
"styleRecalcs": 250,
"styleRecalcDurationMs": 31.75,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 5676.487,
"heapDeltaBytes": -35171724,
"heapUsedBytes": 83925396,
"domNodes": -228,
"jsHeapTotalBytes": 618496,
"scriptDurationMs": 2772.152,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.69901185770751,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "viewport-pan-sweep",
"durationMs": 9291.830000000005,
"styleRecalcs": 251,
"styleRecalcDurationMs": 32.637,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 6097.540999999999,
"heapDeltaBytes": -36098960,
"heapUsedBytes": 82996832,
"domNodes": -228,
"jsHeapTotalBytes": -167936,
"scriptDurationMs": 2966.9529999999995,
"eventListeners": -142,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.665827338129496,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-idle",
"durationMs": 16805.037999999968,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 16373.354,
"heapDeltaBytes": -28388624,
"heapUsedBytes": 196287180,
"domNodes": -3045,
"jsHeapTotalBytes": 8925184,
"scriptDurationMs": 271.87,
"eventListeners": -14915,
"totalBlockingTimeMs": 0,
"frameDurationMs": 29.65784832451499,
"p95FrameDurationMs": 66.59999999999854
},
{
"name": "vue-large-graph-idle",
"durationMs": 12726.081000000022,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 12293.677,
"heapDeltaBytes": -19085172,
"heapUsedBytes": 184963940,
"domNodes": -3045,
"jsHeapTotalBytes": 10141696,
"scriptDurationMs": 222.69,
"eventListeners": -14915,
"totalBlockingTimeMs": 0,
"frameDurationMs": 21.552881355932204,
"p95FrameDurationMs": 33.400000000001455
},
{
"name": "vue-large-graph-pan",
"durationMs": 20682.10799999997,
"styleRecalcs": 144,
"styleRecalcDurationMs": 17.36600000000002,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 20510.766,
"heapDeltaBytes": -37505116,
"heapUsedBytes": 191266776,
"domNodes": -3045,
"jsHeapTotalBytes": 4329472,
"scriptDurationMs": 1073.347,
"eventListeners": -14915,
"totalBlockingTimeMs": 73,
"frameDurationMs": 40.395703125,
"p95FrameDurationMs": 83.30000000000291
},
{
"name": "vue-large-graph-pan",
"durationMs": 20817.677000000003,
"styleRecalcs": 153,
"styleRecalcDurationMs": 18.197999999999993,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 20588.596,
"heapDeltaBytes": -11907596,
"heapUsedBytes": 192855312,
"domNodes": -3045,
"jsHeapTotalBytes": 11145216,
"scriptDurationMs": 925.9560000000001,
"eventListeners": -14915,
"totalBlockingTimeMs": 170,
"frameDurationMs": 41.10513833992094,
"p95FrameDurationMs": 83.39999999999782
},
{
"name": "vue-node-resize-workload",
"durationMs": 6102.001000000087,
"styleRecalcs": 20,
"styleRecalcDurationMs": 81.266,
"layouts": 20,
"layoutDurationMs": 23.733000000000004,
"taskDurationMs": 5998.697,
"heapDeltaBytes": 7804532,
"heapUsedBytes": 205807732,
"domNodes": 0,
"jsHeapTotalBytes": 1310720,
"scriptDurationMs": 109.443,
"eventListeners": 0,
"totalBlockingTimeMs": 453,
"frameDurationMs": 34.95085714285715,
"p95FrameDurationMs": 66.70000000000073
},
{
"name": "vue-node-resize-workload",
"durationMs": 6181.5279999999575,
"styleRecalcs": 20,
"styleRecalcDurationMs": 83.11000000000001,
"layouts": 20,
"layoutDurationMs": 24.455999999999992,
"taskDurationMs": 6129.505999999999,
"heapDeltaBytes": -5160816,
"heapUsedBytes": 224039960,
"domNodes": 0,
"jsHeapTotalBytes": 786432,
"scriptDurationMs": 114.416,
"eventListeners": 0,
"totalBlockingTimeMs": 454,
"frameDurationMs": 36.58698224852071,
"p95FrameDurationMs": 66.70000000000073
},
{
"name": "vue-zoom-out-idle",
"durationMs": 6468.993999999952,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 6125.785000000003,
"heapDeltaBytes": -26145068,
"heapUsedBytes": 167204776,
"domNodes": -3045,
"jsHeapTotalBytes": 15159296,
"scriptDurationMs": 113.61499999999997,
"eventListeners": -14915,
"totalBlockingTimeMs": 0,
"frameDurationMs": 21.994897959183675,
"p95FrameDurationMs": 33.400000000001455
},
{
"name": "vue-zoom-out-idle",
"durationMs": 6556.4789999999675,
"styleRecalcs": 0,
"styleRecalcDurationMs": 0,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 6358.021000000001,
"heapDeltaBytes": 3520192,
"heapUsedBytes": 176113452,
"domNodes": 0,
"jsHeapTotalBytes": -40108032,
"scriptDurationMs": 110.35700000000004,
"eventListeners": 0,
"totalBlockingTimeMs": 0,
"frameDurationMs": 22.56494845360824,
"p95FrameDurationMs": 33.400000000001455
},
{
"name": "workflow-execution",
"durationMs": 537.4920000000429,
"styleRecalcs": 21,
"styleRecalcDurationMs": 29.881999999999998,
"layouts": 5,
"layoutDurationMs": 2.388,
"taskDurationMs": 267.09,
"heapDeltaBytes": -41906100,
"heapUsedBytes": 54690304,
"domNodes": -119,
"jsHeapTotalBytes": 4194304,
"scriptDurationMs": 18.702,
"eventListeners": -62,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "workflow-execution",
"durationMs": 521.302999999989,
"styleRecalcs": 16,
"styleRecalcDurationMs": 23.099999999999998,
"layouts": 4,
"layoutDurationMs": 2.281,
"taskDurationMs": 237.96800000000002,
"heapDeltaBytes": -41860428,
"heapUsedBytes": 54566012,
"domNodes": -141,
"jsHeapTotalBytes": 2883584,
"scriptDurationMs": 12.233,
"eventListeners": -52,
"totalBlockingTimeMs": 0,
"frameDurationMs": 17.81379310344826,
"p95FrameDurationMs": 33.19999999999982
},
{
"name": "agent-composer-empty-conversation-typing",
"durationMs": 997.7059999999938,
"styleRecalcs": 78,
"styleRecalcDurationMs": 108.806,
"layouts": 76,
"layoutDurationMs": 23.588,
"taskDurationMs": 945.36,
"heapDeltaBytes": 18629648,
"heapUsedBytes": 87649632,
"domNodes": 76,
"jsHeapTotalBytes": 262144,
"scriptDurationMs": 390.82900000000006,
"eventListeners": 876,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.664999999999996,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "agent-composer-empty-conversation-typing",
"durationMs": 966.635999999994,
"styleRecalcs": 78,
"styleRecalcDurationMs": 110.08900000000001,
"layouts": 76,
"layoutDurationMs": 24.118000000000002,
"taskDurationMs": 936.3509999999999,
"heapDeltaBytes": 18765456,
"heapUsedBytes": 90360760,
"domNodes": 76,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 394.08099999999996,
"eventListeners": 876,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66610169491524,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "agent-composer-long-conversation-typing",
"durationMs": 1406.714000000079,
"styleRecalcs": 78,
"styleRecalcDurationMs": 246.816,
"layouts": 76,
"layoutDurationMs": 25.112,
"taskDurationMs": 1380.212,
"heapDeltaBytes": 8786844,
"heapUsedBytes": 117710624,
"domNodes": 76,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 472.906,
"eventListeners": 876,
"totalBlockingTimeMs": 0,
"frameDurationMs": 17.49875000000002,
"p95FrameDurationMs": 16.80000000000109
},
{
"name": "agent-composer-long-conversation-typing",
"durationMs": 1415.6100000000151,
"styleRecalcs": 78,
"styleRecalcDurationMs": 250.14799999999997,
"layouts": 76,
"layoutDurationMs": 25.182,
"taskDurationMs": 1392.669,
"heapDeltaBytes": 8274552,
"heapUsedBytes": 117242416,
"domNodes": 76,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 476.681,
"eventListeners": 876,
"totalBlockingTimeMs": 0,
"frameDurationMs": 18.162820512820524,
"p95FrameDurationMs": 33.29999999999927
}
]
}🔬 E2E Coverage
Top 10 files by uncovered lines
🧩 Custom Nodes Ecosystem Matrix: ✅ PassedExecutes the frontend JS of every registry pack that ships any, against this change, in both renderers. The eight
On a pin-bump PR, a red gate means the ecosystem moved — not that your diff broke something. That is what the bump PR is for. Outside a pin bump or a cold rebuild of the unpinned tail, a red means the diff. Note the corpus size and the executed count are different numbers, and the verdict is against the executed count: roughly 3,100 of ~5,100 packs ship no extension-shaped JS and are skipped. Reproduce locally# build one shard's fixture, then run it
python3 scripts/registry-census/build_matrix.py --shard 1/4
MATRIX_RENDERER=vue MATRIX_VUE=1 \
npx vitest run --config vitest.matrix.config.mts
Full docs: scripts/registry-census/README.md |
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Repository: Comfy-Org/ComfyUI_frontend/.coderabbit.yaml Review profile: ASSERTIVE Plan: Team Run ID: 📒 Files selected for processing (6)
Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 5 remain after this review. Your free on-demand review promotion remains active until October 9, 2026 at 6:00 PM UTC. 📝 WalkthroughWalkthroughThe agent composer records starter-prompt identity and click metadata when a suggestion chip is selected. It carries the prompt source through draft submission and includes prompt identifiers in message-sent telemetry. Telemetry types, registry dispatch, and cloud and host providers support the click event. ChangesAgent starter-prompt attribution
Priority: ⚪ Not assessed Estimated code review effort: 3 (Moderate) | ~25 minutes Change: Feature Sequence Diagram(s)sequenceDiagram
actor User
participant EmptyState
participant AgentPanel
participant Composer
participant useComposer
participant TelemetryRegistry
participant useAgentDraftSubmission
participant agentComposerStore
participant AgentPanelRoot
User->>EmptyState: Click starter-prompt chip
EmptyState->>AgentPanel: Emit text and prompt attribution
AgentPanel->>Composer: Insert text and attribution
Composer->>useComposer: Call insert with attribution
useComposer->>TelemetryRegistry: Track prompt-click metadata
User->>useAgentDraftSubmission: Submit draft
useAgentDraftSubmission->>agentComposerStore: Capture starter-prompt source
useAgentDraftSubmission->>AgentPanelRoot: Pass submission metadata
AgentPanelRoot->>TelemetryRegistry: Track message-sent metadata
Suggested reviewers: Merge Risk: 🔵 Low · up to Replacing a starter prompt with a fully typed message can leave the old chip attribution on the send event. This affects analytics rather than message delivery, so the merge risk is bounded. 🚥 Pre-merge checks | ✅ 7✅ Passed checks (7 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
Codecov Report✅ All modified and coverable lines are covered by tests. @@ Coverage Diff @@
## main #18991 +/- ##
==========================================
- Coverage 85.10% 85.01% -0.10%
==========================================
Files 2406 2337 -69
Lines 135477 133470 -2007
Branches 36591 34038 -2553
==========================================
- Hits 115296 113464 -1832
+ Misses 19713 19643 -70
+ Partials 468 363 -105
Flags with carried forward coverage won't be shown. Click here to find out more.
... and 79 files with indirect coverage changes 🚀 New features to boost your workflow:
|
At the panel's default width (PANEL_MIN_WIDTH, 420px) "Find an image-to-video workflow that works on my GPU" needed 375px in a 344px label box, so it ellipsized to "…workflow that works o…" and cut off the GPU — the part doing the work. Kai's shorter line measures 304px and fits, and with it all five local chips render in full at the default width. The slot ids stay exactly as they are. Per Christian, #18991 measures the same five positions across copy and distribution variants, so identity is positional by design and `prompt_text_hash` is what separates one wording from the next. This is a copy variant under a stable slot, which is the case that design was built for.
…19310) ## Summary Shortens the second local/Desktop starter prompt so its full capability hint fits at the panel's default width: ```diff - Find an image-to-video workflow that works on my GPU + An image-to-video workflow that fits my GPU ``` This is a stacked, copy-only follow-up: #19310 → #19283 → #18991. Merge the parent instrumentation PR #18991 first, then the Cloud/local copy PR #19283, then this PR. ## Scope - Changes `src/locales/en/main.json` and the matching `EmptyState.test.ts` fixture only (2 insertions, 2 deletions). - Preserves the five stable positional `AgentStarterPromptId` values from #18991 byte-for-byte: `generate_image`, `list_workflows`, `find_workflow`, `explain_selected_node`, and `build_video_workflow`. - Does not add contextual prompt selection, distribution-scoped IDs, flags, allocation, or telemetry-contract changes. Jo approved the shortened line in the example-prompts Slack thread after the original local copy ellipsized at the default panel width. The rendered measurement reported on this PR is 304 px for the approved line versus 375 px for the prior line in a 344 px label box. ## Verification At exact head `1dcd9740db5c19eeb14643f3ef753fbdfea48102`, the complete GitHub matrix is terminal green: **95 successful checks, 15 intentional skips, and 0 failed, cancelled, queued, or running checks**. The focused local verification published with the change is 326 passing tests across the four affected Agent files, plus clean `typecheck`, `typecheck:browser`, and format checks. The full matrix independently covers unit, browser, ecosystem, lint/typecheck, security, performance, Storybook, and affected web suites. ## Review and merge gate `dante01yoon` is the named human reviewer across this stack. This PR is ready for review, but remains parent-first: #18991 and #19283 must be approved and land before #19310 can merge. Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
|
@ryan-di — rotating the starter-prompt delivery gate after the second named owner was silent for 2h: #18991 is the terminal-green, mergeable root; #19283 is the terminal-green, mergeable child and is now assigned to you too. Please review parent-first: #18991 then #19283. Approved #19310 has already merged into #19283’s feature branch; neither parent is on main yet. |
|
|
|
|
…19310) ## Summary Shortens the second local/Desktop starter prompt so its full capability hint fits at the panel's default width: ```diff - Find an image-to-video workflow that works on my GPU + An image-to-video workflow that fits my GPU ``` This is a stacked, copy-only follow-up: #19310 → #19283 → #18991. Merge the parent instrumentation PR #18991 first, then the Cloud/local copy PR #19283, then this PR. ## Scope - Changes `src/locales/en/main.json` and the matching `EmptyState.test.ts` fixture only (2 insertions, 2 deletions). - Preserves the five stable positional `AgentStarterPromptId` values from #18991 byte-for-byte: `generate_image`, `list_workflows`, `find_workflow`, `explain_selected_node`, and `build_video_workflow`. - Does not add contextual prompt selection, distribution-scoped IDs, flags, allocation, or telemetry-contract changes. Jo approved the shortened line in the example-prompts Slack thread after the original local copy ellipsized at the default panel width. The rendered measurement reported on this PR is 304 px for the approved line versus 375 px for the prior line in a 344 px label box. ## Verification At exact head `1dcd9740db5c19eeb14643f3ef753fbdfea48102`, the complete GitHub matrix is terminal green: **95 successful checks, 15 intentional skips, and 0 failed, cancelled, queued, or running checks**. The focused local verification published with the change is 326 passing tests across the four affected Agent files, plus clean `typecheck`, `typecheck:browser`, and format checks. The full matrix independently covers unit, browser, ecosystem, lint/typecheck, security, performance, Storybook, and affected web suites. ## Review and merge gate `dante01yoon` is the named human reviewer across this stack. This PR is ready for review, but remains parent-first: #18991 and #19283 must be approved and land before #19310 can merge. Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
… they produce (#18991) The five empty-state starter prompts are the cheapest lever on the first-run path and nothing measures them. `input_method: 'suggestion'` on `app:agent_message_sent` already says a chip was used; it cannot say **which** chip, and nothing at all counts a chip that was clicked and never sent. So "is prompt 3 the one that activates people?" is unanswerable, and any experiment on the prompt set would be unreadable. **New: `app:agent_starter_prompt_clicked`** — one event per click, never on render, hover or focus. | property | why it is there | |---|---| | `prompt_id` | The chip's **slot**, from a frozen table (`starterPrompts.ts`), not its display text — the copy is owned by product and is changing | | `prompt_index`, `prompt_count` | A reordered or resized set is visible instead of silently re-labelling slots | | `prompt_text_hash` | FNV-1a, 8 hex, of the displayed text. Without it a copy change under a stable `prompt_id` is indistinguishable from the slot performing differently. Not the text, not reversible | | `locale` | The locale that produced the hash — a translation is a different hash of the same prompt | | `click_id` | Minted per click; the join key to the send | | `draft_was_empty` | Inserting **appends**, so `false` marks a submitted draft that is a mix of this chip and something else | **Extended: `app:agent_message_sent`** gains `starter_prompt_id` and `starter_prompt_click_id`, so click → send → turn is one chain instead of three unrelated counts. Both are `null` for a typed message, which is the other arm of the question. No copy change, no change to the prompt set, no change to submission semantics, no flag or config touched. The click still inserts and still does not send (`browser_tests/tests/agent/agentPanel.spec.ts` covers that and is untouched). - **Attribution lives and dies with `promptOrigin`**: held with the submission snapshot, cleared at `startSubmission`, restored by `takeFailedSubmission` so a retry is still that chip's message, dropped when the conversation's edit action replaces the prompt. Two sources of truth that could disagree about "did this come from a chip" would be worse than one, so they share a lifecycle. Invariant, checkable in the data: `starter_prompt_id` is non-null exactly when `input_method === 'suggestion'`. - **Reporting is gated on the caller identifying a prompt.** `insert()` without an attribution still inserts and is still recorded as `suggestion`, but reports no click — so a future affordance that reuses `insert()` cannot inflate the starter-prompt counts. - **A sixth prompt with no id reports `unregistered`** rather than borrowing its neighbour's identity, and `starterPrompts.test.ts` fails with a comment telling you to add the id. Graceful in data, loud in CI. - **The last chip clicked owns the send**, because inserting appends and a send is one message. `draft_was_empty: false` is how the mixed case is found rather than assumed away. `pnpm vitest run src/workbench/extensions/agent src/platform/telemetry` — 177 files, 2918 passing. New coverage, one case per branch: - `utils/starterPrompts.test.ts` — id table length matches the locale array; out-of-table slot → `unregistered`; id survives a copy change while the hash does not; hash shape and determinism. - `components/agent/EmptyState.test.ts` — table-driven over all five chips (slot id and index); nothing emitted before a click; two clicks are two events; the payload carries the hash and **not** the text. - `composables/agent/useComposer.test.ts` — one click event per identified insert, with the click id the send will carry; second click reports `draft_was_empty: false`; unidentified insert reports nothing and leaves no stale attribution. - `stores/agent/agentComposerStore.test.ts` — chip survives a reword, cleared at submission, restored for a failed retry, dropped by the edit action. - `composables/agent/useAgentDraftSubmission.test.ts` — the chip reaches `SubmissionMeta`, is not carried into the next send, and is restored for the retry with a fresh `clientMessageId`. - `AgentPanelRoot.test.ts` — the full chain end to end: the click event's `click_id` is the send event's `starter_prompt_click_id`. - Registry / PostHog / host-sink conformance cases for the new event. Not on merge. Production serves a pinned frontend SHA (`08abb06e41` as of 2026-09-26, with `canary.enabled: false`), and `app:agent_message_sent`'s existing `input_method` is a live example of the gap: it is in `main`, absent from the served build, and has **0 events on `cloud.comfy.org`** against 103 on `agent.comfy.org` in 30 days. This event needs merge → `cloud/1.54` backport → pin bump → argocd sync before a production user can emit it. Any query on it must split by `$host`.
…nd sends (#18991) (#19488) Backports #18991 to `cloud/1.54`. This carries starter-prompt click telemetry and correlates the subsequent send with prompt/click IDs while preserving the release branch Agent and telemetry APIs. It is based on the already-merged consent telemetry predecessor #19469. Conflict resolution: - preserved the release-branch Host telemetry and Agent panel test structure - added the starter-prompt event/method and send attribution fields - retained release-branch component behavior while carrying the runtime attribution path and focused coverage Validation: - pre-commit formatting, lint, ESLint, and `typecheck:app` passed - 9 focused test files passed: 326 tests - focused AgentPanelRoot attribution/workflow tests passed: 5 tests - full AgentPanelRoot file also exposed five unrelated asset/Response harness failures; the three attribution expectation failures were updated and pass in the focused run Original: #18991
…d sends (#18991) (#19490) ## Summary Backports #18991 so core 1.54 records each starter-prompt click and attributes the resulting message send to the same click. ## Changes - **What**: Adds stable prompt IDs, copy hashes, per-click IDs, click telemetry, and send attribution with focused unit coverage. - **Dependencies**: `core/1.54` did not contain #18908 (the 1.54 activation-funnel telemetry substrate already present on `cloud/1.54`), so this backport includes that prerequisite while preserving the newer consent-outcome telemetry from #19470. ## Review Focus Confirm the activation-funnel prerequisite is appropriate for core 1.54 and the target-specific `sendMessage` integration preserves the four-argument core API. ## Verification - `pnpm vitest run` on the eight focused starter-prompt, composer-store, submission, registry, PostHog, and host-sink suites: 203 passed - Focused `AgentPanelRoot` click-to-send chain: 1 passed - `pnpm typecheck`: passed - Pre-commit lint, format, style, and typecheck hooks: passed - Pre-push `knip`: passed --------- Co-authored-by: Hunter <huntcsg@users.noreply.github.com>
…d sends (#18991) (#19489) Backports #18991 to `core/1.55`. This preserves the release branch APIs while adding stable starter-prompt IDs, click telemetry, and click-to-send attribution (including retry restoration). The release branch predates some mainline composer plumbing, so the attribution lifecycle is adapted directly rather than importing unrelated newer panel code. Verification: - `pnpm vitest run src/workbench/extensions/agent/composables/agent/useComposer.test.ts src/workbench/extensions/agent/utils/starterPrompts.test.ts src/platform/telemetry/TelemetryRegistry.test.ts src/platform/telemetry/providers/cloud/PostHogTelemetryProvider.test.ts src/platform/telemetry/providers/host/HostTelemetrySink.test.ts` (153 passed) - `pnpm typecheck` (passed) - pre-push `knip --cache` (passed)
…nd sends (#18991) (#19491) Backports #18991 to cloud/1.55 after the automatic backport conflicted. Preserves the 1.55 Agent APIs while carrying stable starter-prompt IDs, click telemetry, click-to-send attribution, input method, and retry restoration. Validation: 153 focused tests passed. Source merge: d7d6752
## Summary - give Cloud prompts focused on ready-to-run creation, video, workflow discovery, and repair - give local/Desktop prompts focused on local GPU workflows, workflow repair, and missing-node setup - apply Jo's exact five Cloud and five local prompt strings from the [example-prompts thread](https://comfy-organization.slack.com/archives/C0A8Z4U7Y1K/p1790573450258129) - preserve the five stable telemetry slots from #18991 so copy changes remain measurable This is stacked on #18991 (starter-prompt click and send attribution); merge that PR first. ## Verification Exact head: `acf46b77f9e7f6242dd5d2bb25eaeac66dca1bf6` - application unit suite: 1,707 files; 24,337 passed, 32 expected-fail, 7 skipped (24,376 total) — [run](https://github.com/Comfy-Org/ComfyUI_frontend/actions/runs/36487711019) - Cloud Playwright: 454 tests; 452 passed, 1 skipped, 1 unrelated flaky retry in `agentRateLimitDuplicateSubmit.spec.ts` — [job](https://github.com/Comfy-Org/ComfyUI_frontend/actions/runs/36487711311/job/109152438679) - full rollup: 95 successful checks, 14 intentional skips, 0 failed, cancelled, or pending - focused local verification: 4 files, 326 passed; application/browser typecheck, ESLint, formatting, Knip, Fallow, and `git diff --check` all passed The ten prompt strings match Jo's approved Slack copy byte-for-byte. The stable IDs remain `generate_image`, `list_workflows`, `find_workflow`, `explain_selected_node`, and `build_video_workflow`. ## Delivery gate #18991 is terminal green and assigned to @dante01yoon, but still needs human approval and must land before this stacked PR can merge. --------- Co-authored-by: jojodecayz <121620462+jojodecayz@users.noreply.github.com> Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
… local (#19283) (#19569) ## Summary Backports #19283 to `core/1.55`. - ships the approved five-prompt distributions for Cloud and local/Desktop - preserves the stable telemetry slot IDs introduced by #18991 - retains core/1.55-specific test structure while carrying the source feature and focused coverage ## Conflict resolution The release branch already contains the manually backported #18991 implementation, but its test scaffolding differs from `main`. Conflicted test-only hunks were resolved against the release line, then the slot-ID assertions needed by this feature were applied directly. Production prompt selection, copy, localization, and telemetry types are preserved from the source PR. ## Verification - `pnpm typecheck:app` - focused Vitest: 7 files, 270 tests passed - pre-push `knip --cache` - `git diff --check` ## Source - Source PR: #19283 - Dependency/backport context: #19489 Co-authored-by: jojodecayz <121620462+jojodecayz@users.noreply.github.com> Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
|
|
|
|
Why
The five empty-state starter prompts are the cheapest lever on the first-run path and nothing measures them.
input_method: 'suggestion'onapp:agent_message_sentalready says a chip was used; it cannot say which chip, and nothing at all counts a chip that was clicked and never sent. So "is prompt 3 the one that activates people?" is unanswerable, and any experiment on the prompt set would be unreadable.What
New:
app:agent_starter_prompt_clicked— one event per click, never on render, hover or focus.prompt_idstarterPrompts.ts), not its display text — the copy is owned by product and is changingprompt_index,prompt_countprompt_text_hashprompt_idis indistinguishable from the slot performing differently. Not the text, not reversiblelocaleclick_iddraft_was_emptyfalsemarks a submitted draft that is a mix of this chip and something elseExtended:
app:agent_message_sentgainsstarter_prompt_idandstarter_prompt_click_id, so click → send → turn is one chain instead of three unrelated counts. Both arenullfor a typed message, which is the other arm of the question.Behaviour: unchanged
No copy change, no change to the prompt set, no change to submission semantics, no flag or config touched. The click still inserts and still does not send (
browser_tests/tests/agent/agentPanel.spec.tscovers that and is untouched).Design notes
promptOrigin: held with the submission snapshot, cleared atstartSubmission, restored bytakeFailedSubmissionso a retry is still that chip's message, dropped when the conversation's edit action replaces the prompt. Two sources of truth that could disagree about "did this come from a chip" would be worse than one, so they share a lifecycle. Invariant, checkable in the data:starter_prompt_idis non-null exactly wheninput_method === 'suggestion'.insert()without an attribution still inserts and is still recorded assuggestion, but reports no click — so a future affordance that reusesinsert()cannot inflate the starter-prompt counts.unregisteredrather than borrowing its neighbour's identity, andstarterPrompts.test.tsfails with a comment telling you to add the id. Graceful in data, loud in CI.draft_was_empty: falseis how the mixed case is found rather than assumed away.Tests
pnpm vitest run src/workbench/extensions/agent src/platform/telemetry— 177 files, 2918 passing. New coverage, one case per branch:utils/starterPrompts.test.ts— id table length matches the locale array; out-of-table slot →unregistered; id survives a copy change while the hash does not; hash shape and determinism.components/agent/EmptyState.test.ts— table-driven over all five chips (slot id and index); nothing emitted before a click; two clicks are two events; the payload carries the hash and not the text.composables/agent/useComposer.test.ts— one click event per identified insert, with the click id the send will carry; second click reportsdraft_was_empty: false; unidentified insert reports nothing and leaves no stale attribution.stores/agent/agentComposerStore.test.ts— chip survives a reword, cleared at submission, restored for a failed retry, dropped by the edit action.composables/agent/useAgentDraftSubmission.test.ts— the chip reachesSubmissionMeta, is not carried into the next send, and is restored for the retry with a freshclientMessageId.AgentPanelRoot.test.ts— the full chain end to end: the click event'sclick_idis the send event'sstarter_prompt_click_id.When this can be read in production
Not on merge. Production serves a pinned frontend SHA (
08abb06e41as of 2026-09-26, withcanary.enabled: false), andapp:agent_message_sent's existinginput_methodis a live example of the gap: it is inmain, absent from the served build, and has 0 events oncloud.comfy.orgagainst 103 onagent.comfy.orgin 30 days. This event needs merge →cloud/1.54backport → pin bump → argocd sync before a production user can emit it. Any query on it must split by$host.