[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7DNBSKN21HEGA31ZZNGHXJ6":3,"skill-same:01M7DNBSKN21HEGA31ZZNGHXJ6":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M7DNBSKN21HEGA31ZZNGHXJ6","chat-perf",null,"Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.","DragonBallerZ\u002Fdragon-ide","DragonBallerZ","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F199459338?v=4","01M7DN83GSDPNQH3ACEV21SKTQ","Z McDragon","01M7DN83HDR4KMJAMTCB5S4DCJ",12,"MIT","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fchat-perf\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: chat-perf\ndescription: Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.\n---\n\n# Chat Performance Testing\n\n## When to use\n\n- Before\u002Fafter modifying chat rendering code (`chatListRenderer.ts`, `chatInputPart.ts`, markdown rendering)\n- When changing the streaming response pipeline or SSE processing\n- When modifying disposable\u002Flifecycle patterns in chat components\n- To compare performance between two VS Code releases\n- In CI to gate PRs that touch chat UI code\n\n## Quick start\n\n```bash\n# Run perf regression test (compares local dev build vs VS Code 1.115.0):\nnpm run perf:chat -- --scenario text-only --runs 3\n\n# Run all scenarios with no baseline (just measure):\nnpm run perf:chat -- --no-baseline --runs 3\n\n# Compare two local builds (apples-to-apples):\nnpm run perf:chat -- --build \u002Fpath\u002Fto\u002Fbuild-A --baseline-build \u002Fpath\u002Fto\u002Fbuild-B --runs 5\n\n# Build a local production package and compare against a release:\nnpm run perf:chat -- --production-build --baseline-build 1.115.0 --runs 5\n\n# Run memory leak check (10 messages in one session):\nnpm run perf:chat-leak\n\n# Run leak check with more messages for accuracy:\nnpm run perf:chat-leak -- --messages 20 --verbose\n```\n\n## Perf regression test\n\n**Script:** `scripts\u002Fchat-simulation\u002Ftest-chat-perf-regression.js` **npm:** `npm run perf:chat`\n\nLaunches VS Code via Playwright Electron, opens the chat panel, sends a message with a mock LLM response, and measures timing, layout, and rendering metrics. By default, downloads VS Code 1.115.0 as a baseline, benchmarks it, then benchmarks the local dev build and compares.\n\n> **You don't always need a baseline.** A baseline exists only for *comparison* (regression detection). If you just want the current build's numbers — profiling a single change, capturing traces\u002Fheap snapshots, or iterating on a scenario — pass `--no-baseline` to skip downloading and benchmarking the baseline entirely (roughly halves runtime). Baseline comparison is what turns raw measurements into a pass\u002Ffail verdict; without it you still get all the metrics, just no verdict.\n\n### Key flags\n\n| Flag | Default | Description |\n|---|---|---|\n| `--runs \u003Cn>` | `5` | Runs per scenario. More = more stable. Use 5+ for CI. |\n| `--scenario \u003Cid>` \u002F `-s` | all | Scenario to test (repeatable). See `common\u002Fperf-scenarios.js`. |\n| `--build \u003Cpath\\|ver>` \u002F `-b` | local dev | Build to test. Accepts path or version (`1.110.0`, `insiders`, commit hash). |\n| `--baseline \u003Cpath>` | — | Compare against a previously saved baseline JSON file. |\n| `--baseline-build \u003Cpath\\|ver>` | `1.115.0` | Version or local path to benchmark as baseline. |\n| `--no-baseline` | — | Skip the baseline entirely — just measure the test build (no download, no comparison, ~2× faster). Use when you only need raw numbers, not a regression verdict. |\n| `--save-baseline` | — | Save results as the new baseline (requires `--baseline \u003Cpath>`). |\n| `--resume \u003Cpath>` | — | Resume a previous run, adding more iterations to increase confidence. |\n| `--threshold \u003Cfrac>` | `0.2` | Regression threshold (0.2 = flag if 20% slower). |\n| `--production-build` | — | Build a local bundled package via `gulp vscode` for comparison against a release baseline. |\n| `--no-cache` | — | Ignore cached baseline data, always run fresh. |\n| `--force` | — | Skip build mode mismatch confirmation prompt. |\n| `--ci` | — | CI mode: write Markdown summary to `ci-summary.md` (implies `--no-cache`, `--heap-snapshots`, `--cleanup-diagnostics`). |\n| `--heap-snapshots` | — | Take heap snapshots after each run (slow; auto-enabled in `--ci` mode). |\n| `--gc-object-stats` | — | **GC deep-dives only.** Enables V8 `gc_stats` tracing (per-type heap object dump on every GC). ⚠️ Corrupts all timing metrics — a major GC landing mid-request adds ~550ms — so never use it for benchmarking. Off by default; prefer heap snapshots for memory analysis. |\n| `--cleanup-diagnostics` | — | Delete heap snapshots, CPU profiles, and traces to save disk. During runs, only the latest run's files are kept; after comparison, files for non-regressed scenarios are deleted. Auto-enabled in `--ci` mode. |\n| `--setting \u003Ck=v>` | — | Set a VS Code setting override for all builds (repeatable). |\n| `--test-setting \u003Ck=v>` | — | Set a VS Code setting override for the test build only. |\n| `--baseline-setting \u003Ck=v>` | — | Set a VS Code setting override for the baseline build only. |\n| `--verbose` | — | Print per-run details including response content. |\n\n### Comparing two remote builds\n\n```bash\n# Compare 1.110.0 against 1.115.0 (no local build needed):\nnpm run perf:chat -- --build 1.110.0 --baseline-build 1.115.0 --runs 5\n```\n\n### Comparing two local builds\n\nBoth `--build` and `--baseline-build` accept local paths to VS Code executables. This enables apples-to-apples comparisons between any two builds:\n\n```bash\n# Compare two dev builds (e.g. feature branch vs main):\nnpm run perf:chat -- \\\n  --build .build\u002Felectron\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --baseline-build \u002Fpath\u002Fto\u002Fother\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --runs 5\n\n# Compare two production builds:\nnpm run perf:chat -- \\\n  --build ..\u002FVSCode-darwin-arm64-feature\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --baseline-build ..\u002FVSCode-darwin-arm64-main\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --runs 5\n```\n\nLocal path baselines are never cached (the build may change between runs). Version string baselines are cached for reuse.\n\n### Build modes and mismatch detection\n\nThe tool classifies builds into three modes based on the executable path:\n\n| Mode | Source | Characteristics |\n|---|---|---|\n| `dev` | `.build\u002Felectron\u002F` (local dev) | Unbundled sources, `VSCODE_DEV=1`, `NODE_ENV=development`. Higher memory and startup overhead. |\n| `production` | `..\u002FVSCode-\u003Cplatform>-\u003Carch>\u002F` (from `gulp vscode`) | Bundled JS, no dev flags. Matches release characteristics but uses local source. |\n| `release` | `.vscode-test\u002F` (downloaded via `@vscode\u002Ftest-electron`) | Official published build. |\n\nWhen test and baseline builds have different modes (e.g. dev vs release), the tool shows a warning and prompts for confirmation. Use `--force` or `--ci` to skip the prompt.\n\nUsing `--production-build` builds a local bundled package via `gulp vscode` for fair comparison against a release baseline. This eliminates dev-mode overhead while still testing your local changes.\n\n```bash\n# Production build vs release baseline (fair comparison):\nnpm run perf:chat -- --production-build --baseline-build 1.115.0 --runs 5\n```\n\n### Settings overrides\n\nUse `--setting`, `--test-setting`, and `--baseline-setting` to inject VS Code settings into the launched instance. This is useful for A\u002FB testing experimental features:\n\n```bash\n# Enable a feature for the test build only:\nnpm run perf:chat -- --test-setting chat.experimental.incrementalRendering.enabled=true --runs 3\n\n# Compare two builds with different settings:\nnpm run perf:chat -- \\\n  --baseline-build \"..\u002Fvscode2\u002F.build\u002Felectron\u002FCode - OSS.app\u002FContents\u002FMacOS\u002FCode - OSS\" \\\n  --baseline-setting chat.experimental.incrementalRendering.enabled=true \\\n  --test-setting chat.experimental.incrementalRendering.enabled=false \\\n  --runs 3\n\n# Set a value for both builds:\nnpm run perf:chat -- --setting chat.mcp.enabled=false --runs 3\n```\n\nPrecedence: `--test-setting` \u002F `--baseline-setting` override `--setting` for the same key. Values are auto-parsed: `true`\u002F`false` become booleans, numbers become numbers, everything else stays a string.\n\n### Resuming a run for more confidence\n\nWhen results exceed the threshold but aren't statistically significant, the tool prints a `--resume` hint. Use it to add more iterations to an existing run:\n\n```bash\n# Initial run with 3 iterations — may be inconclusive:\nnpm run perf:chat -- --scenario text-only --runs 3\n\n# Add 3 more runs to the same results file (both test + baseline):\nnpm run perf:chat -- --resume .chat-simulation-data\u002F2026-04-14T02-15-14\u002Fresults.json --runs 3\n\n# Keep adding until confidence is reached:\nnpm run perf:chat -- --resume .chat-simulation-data\u002F2026-04-14T02-15-14\u002Fresults.json --runs 5\n```\n\n`--resume` loads the previous `results.json` and its associated `baseline-*.json`, runs N more iterations for both builds, merges rawRuns, recomputes stats, and re-runs the comparison. The updated files are written back in-place. You can resume multiple times — samples accumulate.\n\n### Statistical significance\n\nRegression detection uses **Welch's t-test** to avoid false positives from noisy measurements. A metric is only flagged as `REGRESSION` when it both exceeds the threshold AND is statistically significant (p \u003C 0.05). Otherwise it's reported as `(likely noise — p=X, not significant)`.\n\nWith typical variance (cv ≈ 20%), you need:\n- **n ≥ 5** per build to detect a 35% regression at 95% confidence\n- **n ≥ 10** per build to detect a 20% regression reliably\n\nConfidence levels reported: `high` (p \u003C 0.01), `medium` (p \u003C 0.05), `low` (p \u003C 0.1), `none`.\n\n### Exit codes\n\n- `0` — all metrics within threshold, or exceeding threshold but not statistically significant\n- `1` — statistically significant regression detected, or all runs failed\n\n### Scenarios\n\nScenarios are defined in `scripts\u002Fchat-simulation\u002Fcommon\u002Fperf-scenarios.js` and registered via `registerPerfScenarios()`. There are three categories:\n\n- **Content-only** — plain streaming responses (e.g. `text-only`, `large-codeblock`, `rapid-stream`)\n- **Tool-call** — multi-turn scenarios with tool invocations (e.g. `tool-read-file`, `tool-edit-file`)\n- **Multi-turn user** — multi-turn conversations with user follow-ups, thinking blocks (e.g. `thinking-response`, `multi-turn-user`, `long-conversation`)\n\nRun `npm run perf:chat -- --help` to see the full list of registered scenario IDs.\n\n### Metrics collected\n\n- **Timing:** time to first token, time to complete, time to render complete (includes typewriter animation)\n- **Rendering:** layout count, layout duration (ms), style recalculation count, forced reflows, long tasks (>50ms), long animation frame count and duration\n- **Memory:** heap before\u002Fafter, heap delta post-GC (informational, noisy for single requests)\n- **Extension host:** heap before\u002Fafter\u002Fdelta via CDP inspector\n\n### Regression triggers vs informational metrics\n\nOnly these metrics trigger a regression failure (when they exceed the threshold with statistical significance):\n- `timeToFirstToken`, `timeToComplete` — user-perceived latency\n- `layoutDurationMs` — total layout time from the trace (the *real* layout cost)\n- `forcedReflowCount` — forced synchronous layouts are always bad\n- `longTaskCount` — main thread jank\n\nThese are reported but **informational only** (won't fail CI):\n- `longAnimationFrameCount` — number of long animation frames; noisy and compositor-driven, so tracked for signal but does not gate. Real jank is gated via `longTaskCount` \u002F `layoutDurationMs`.\n- `layoutCount` — number of layout ops; inflated by CSS animations (compositor-driven, cheap). A build can do *more but cheaper* layouts, so gate on `layoutDurationMs`, not this count.\n- `recalcStyleCount` — number of style recalcs; inflated by CSS animations (compositor-driven, cheap)\n- `timeToRenderComplete` — includes typewriter animation tail\n- Memory\u002Fheap metrics — too noisy for single-request benchmarks\n\n### Statistics\n\nResults use **IQR-based outlier removal** and **median** (not mean) to handle startup jitter. The **coefficient of variation (cv)** is reported — under 15% is stable, over 15% gets a ⚠ warning. Baseline comparison uses **Welch's t-test** on raw run values to determine statistical significance before flagging regressions. Use 5+ runs to get stable results.\n\n## Memory leak check\n\n**Script:** `scripts\u002Fchat-simulation\u002Ftest-chat-mem-leaks.js` **npm:** `npm run perf:chat-leak`\n\nLaunches one VS Code session, sends N messages sequentially, forces GC between each, and measures renderer heap and DOM node count. Uses **linear regression** on the samples to compute per-message growth rate, which is compared against a threshold.\n\n### Key flags\n\n| Flag | Default | Description |\n|---|---|---|\n| `--messages \u003Cn>` \u002F `-n` | `10` | Number of messages to send. More = more accurate slope. |\n| `--build \u003Cpath\\|ver>` \u002F `-b` | local dev | Build to test. |\n| `--threshold \u003CMB>` | `2` | Max per-message heap growth in MB. |\n| `--setting \u003Ck=v>` | — | Set a VS Code setting override (repeatable). |\n| `--verbose` | — | Print per-message heap\u002FDOM counts. |\n\n### What it measures\n\n- **Heap growth slope** (MB\u002Fmessage) — linear regression over forced-GC heap samples. A leak shows as sustained positive slope.\n- **DOM node growth** (nodes\u002Fmessage) — catches rendering leaks where elements aren't cleaned up. Healthy chat virtualizes old messages so node count plateaus.\n\n### Interpreting results\n\n- `0.3–1.0 MB\u002Fmsg` — normal (V8 internal overhead, string interning)\n- `>2.0 MB\u002Fmsg` — likely leak, investigate retained objects\n- DOM nodes stable after first message — normal (chat list virtualization working)\n- DOM nodes growing linearly — rendering leak, check disposable cleanup\n\n## CI runs & pinpointing regressions\n\nThe perf + leak checks run in CI via the **`.github\u002Fworkflows\u002Fchat-perf.yml`** workflow (a scheduled daily `workflow_dispatch`, plus manual dispatch). Each run benchmarks the current `main` as the **test** build against a fixed release **baseline** (from `config.jsonc`, e.g. `1.122.0`). Because the baseline is fixed, the **test** median for a metric across successive daily runs traces `main`'s trajectory over time — that's what lets you bisect *when* something regressed or went flaky.\n\n### Finding and reading historic runs\n\n```bash\n# List recent runs (most recent first) — note the run IDs and dates\ngh run list --workflow chat-perf.yml -R microsoft\u002Fvscode --limit 30 \\\n  --json databaseId,status,conclusion,createdAt,headBranch \\\n  --jq '.[] | [.databaseId, (.conclusion\u002F\u002F.status), .createdAt, .headBranch] | @tsv'\n\n# See a run's per-job results\ngh run view \u003Crun-id> -R microsoft\u002Fvscode --json jobs \\\n  --jq '.jobs[] | [.name, (.conclusion\u002F\u002F.status)] | @tsv'\n\n# Trigger a run manually against any ref\u002Fversion:\ngh workflow run chat-perf.yml -R microsoft\u002Fvscode --ref main \\\n  -f test_build=\u003Cbranch|sha|version> -f baseline_build=\u003Cversion>\n```\n\n### Artifacts per run (and retention — this matters for old runs)\n\n| Artifact | Contents | Retention |\n|---|---|---|\n| `chat-perf-summary` | Unified `ci-summary.md`: verdicts, per-metric medians ±stddev, **per-run raw tables** | 30 days |\n| `perf-results-\u003Cgroup>` | Everything below **plus traces, CPU profiles, heap snapshots** (large) | 30 days |\n| `leak-results` | Leak log + `chat-simulation-leak-results.json` | 30 days |\n| `perf-summary-\u003Cgroup>` | `results.json` (full per-run metrics incl. `rawRuns`) + `baseline-*.json` | **1 day** |\n\n```bash\n# List a run's artifacts + whether they've expired\ngh api repos\u002Fmicrosoft\u002Fvscode\u002Factions\u002Fruns\u002F\u003Crun-id>\u002Fartifacts \\\n  --jq '.artifacts[] | [.name, .expired] | @tsv'\n\n# Download the human-readable summary (best first stop; survives 30 days)\ngh run download \u003Crun-id> -R microsoft\u002Fvscode -n chat-perf-summary\n```\n\n### Pinpointing where a metric regressed \u002F went flaky\n\n1. **Bisect across dates.** Pull `chat-perf-summary\u002Fci-summary.md` from several runs spanning the window. Compare the **test** median (and ±stddev) for the suspect metric+scenario run-to-run — the day it jumps is when `main` changed. A metric that goes *bimodal* (e.g. a raw-run column showing two clusters like `~250 \u002F ~900`) is the flaky signature; a stable shift is a genuine regression.\n2. **Confirm it's real vs. measurement noise.** Check the per-run raw tables (in `ci-summary.md`) — high ±stddev \u002F bimodal values mean the *median* is being pulled around by a few outlier runs, not a uniform slowdown.\n3. **Deep-dive the cause.** Download `perf-results-\u003Cgroup>` (has the `trace.json` per run) for a slow run and inspect what dominates the slow window. Trick that found the `gc_stats` artifact: sum main-thread `RunTask` durations between two `code\u002Fchat\u002F*` marks (e.g. `willCollectInstructions` → `didCollectInstructions`) — if the window is ~0% busy, it's an async wait or a GC pause, not real work; then look at the largest `X`-phase events in that window (`MajorGC`, `Layout`, etc.).\n4. **Reproduce a suspect commit locally** to bisect precisely: `npm run perf:chat -- --build \u003Csha> --baseline-build \u003Cver> --runs 7` (or two commits directly via `--build \u003CshaA> --baseline-build \u003CshaB>`).\n\n> Tip: `perf-summary-*` (the machine-readable `results.json` with `rawRuns`) is deleted after **1 day**, so for older runs rely on `chat-perf-summary` (raw tables, 30 days) or extract `results.json` from `perf-results-*` (also 30 days).\n\n## Architecture\n\n```\nscripts\u002Fchat-simulation\u002F\n├── common\u002F\n│   ├── mock-llm-server.js    # Mock CAPI server matching @vscode\u002Fcopilot-api URL structure\n│   ├── perf-scenarios.js     # Built-in scenario definitions (content, tool-call, multi-turn)\n│   └── utils.js              # Shared: paths, env setup, stats, launch helpers\n├── config.jsonc              # Default config (baseline version, runs, thresholds)\n├── fixtures\u002F                 # TypeScript fixture files used by tool-call scenarios\n├── test-chat-perf-regression.js\n└── test-chat-mem-leaks.js\n```\n\n### Mock server\n\nThe mock LLM server (`common\u002Fmock-llm-server.js`) implements the full CAPI URL structure from `@vscode\u002Fcopilot-api`'s `DomainService`:\n\n- `GET \u002Fmodels` — returns model metadata\n- `POST \u002Fmodels\u002Fsession` — returns `AutoModeAPIResponse` with `available_models` and `session_token`\n- `POST \u002Fmodels\u002Fsession\u002Fintent` — model router\n- `POST \u002Fchat\u002Fcompletions` — SSE streaming response matching the scenario\n- Agent, session, telemetry, and token endpoints\n\nThe copilot extension connects to this server via `IS_SCENARIO_AUTOMATION=1` mode with `overrideCapiUrl` and `overrideProxyUrl` settings. The `vscode-api-tests` extension must be disabled (`--disable-extension=vscode.vscode-api-tests`) because it contributes a duplicate `copilot` vendor that blocks the real extension's language model provider registration.\n\n### Adding a scenario\n\n1. Add a new entry to the appropriate object (`CONTENT_SCENARIOS`, `TOOL_CALL_SCENARIOS`, or `MULTI_TURN_SCENARIOS`) in `common\u002Fperf-scenarios.js` using the `ScenarioBuilder` API from `common\u002Fmock-llm-server.js`\n2. The scenario is auto-registered by `registerPerfScenarios()` — no manual ID list to update\n3. Run: `npm run perf:chat -- --scenario your-new-scenario --runs 1 --no-baseline --verbose`\n\n## Related skills\n\n- **heap-snapshot-analysis** — When a perf regression or leak check identifies high memory growth, use the heap-snapshot-analysis skill to dig deeper. It can parse `.heapsnapshot` files, compare before\u002Fafter snapshots, group object deltas, and trace retainer paths to find what keeps disposed objects alive. The chat-perf leak check measures overall heap slope; heap-snapshot-analysis finds the specific objects responsible.\n- **auto-perf-optimize** — For launching VS Code, driving a scenario, and capturing heap snapshots or CPU profiles automatically before doing low-level analysis.\n","# Chat Performance Testing\n\n## When to use\n\n- Before\u002Fafter modifying chat rendering code (`chatListRenderer.ts`, `chatInputPart.ts`, markdown rendering)\n- When changing the streaming response pipeline or SSE processing\n- When modifying disposable\u002Flifecycle patterns in chat components\n- To compare performance between two VS Code releases\n- In CI to gate PRs that touch chat UI code\n\n## Quick start\n\n```bash\n# Run perf regression test (compares local dev build vs VS Code 1.115.0):\nnpm run perf:chat -- --scenario text-only --runs 3\n\n# Run all scenarios with no baseline (just measure):\nnpm run perf:chat -- --no-baseline --runs 3\n\n# Compare two local builds (apples-to-apples):\nnpm run perf:chat -- --build \u002Fpath\u002Fto\u002Fbuild-A --baseline-build \u002Fpath\u002Fto\u002Fbuild-B --runs 5\n\n# Build a local production package and compare against a release:\nnpm run perf:chat -- --production-build --baseline-build 1.115.0 --runs 5\n\n# Run memory leak check (10 messages in one session):\nnpm run perf:chat-leak\n\n# Run leak check with more messages for accuracy:\nnpm run perf:chat-leak -- --messages 20 --verbose\n```\n\n## Perf regression test\n\n**Script:** `scripts\u002Fchat-simulation\u002Ftest-chat-perf-regression.js` **npm:** `npm run perf:chat`\n\nLaunches VS Code via Playwright Electron, opens the chat panel, sends a message with a mock LLM response, and measures timing, layout, and rendering metrics. By default, downloads VS Code 1.115.0 as a baseline, benchmarks it, then benchmarks the local dev build and compares.\n\n> **You don't always need a baseline.** A baseline exists only for *comparison* (regression detection). If you just want the current build's numbers — profiling a single change, capturing traces\u002Fheap snapshots, or iterating on a scenario — pass `--no-baseline` to skip downloading and benchmarking the baseline entirely (roughly halves runtime). Baseline comparison is what turns raw measurements into a pass\u002Ffail verdict; without it you still get all the metrics, just no verdict.\n\n### Key flags\n\n| Flag | Default | Description |\n|---|---|---|\n| `--runs \u003Cn>` | `5` | Runs per scenario. More = more stable. Use 5+ for CI. |\n| `--scenario \u003Cid>` \u002F `-s` | all | Scenario to test (repeatable). See `common\u002Fperf-scenarios.js`. |\n| `--build \u003Cpath\\|ver>` \u002F `-b` | local dev | Build to test. Accepts path or version (`1.110.0`, `insiders`, commit hash). |\n| `--baseline \u003Cpath>` | — | Compare against a previously saved baseline JSON file. |\n| `--baseline-build \u003Cpath\\|ver>` | `1.115.0` | Version or local path to benchmark as baseline. |\n| `--no-baseline` | — | Skip the baseline entirely — just measure the test build (no download, no comparison, ~2× faster). Use when you only need raw numbers, not a regression verdict. |\n| `--save-baseline` | — | Save results as the new baseline (requires `--baseline \u003Cpath>`). |\n| `--resume \u003Cpath>` | — | Resume a previous run, adding more iterations to increase confidence. |\n| `--threshold \u003Cfrac>` | `0.2` | Regression threshold (0.2 = flag if 20% slower). |\n| `--production-build` | — | Build a local bundled package via `gulp vscode` for comparison against a release baseline. |\n| `--no-cache` | — | Ignore cached baseline data, always run fresh. |\n| `--force` | — | Skip build mode mismatch confirmation prompt. |\n| `--ci` | — | CI mode: write Markdown summary to `ci-summary.md` (implies `--no-cache`, `--heap-snapshots`, `--cleanup-diagnostics`). |\n| `--heap-snapshots` | — | Take heap snapshots after each run (slow; auto-enabled in `--ci` mode). |\n| `--gc-object-stats` | — | **GC deep-dives only.** Enables V8 `gc_stats` tracing (per-type heap object dump on every GC). ⚠️ Corrupts all timing metrics — a major GC landing mid-request adds ~550ms — so never use it for benchmarking. Off by default; prefer heap snapshots for memory analysis. |\n| `--cleanup-diagnostics` | — | Delete heap snapshots, CPU profiles, and traces to save disk. During runs, only the latest run's files are kept; after comparison, files for non-regressed scenarios are deleted. Auto-enabled in `--ci` mode. |\n| `--setting \u003Ck=v>` | — | Set a VS Code setting override for all builds (repeatable). |\n| `--test-setting \u003Ck=v>` | — | Set a VS Code setting override for the test build only. |\n| `--baseline-setting \u003Ck=v>` | — | Set a VS Code setting override for the baseline build only. |\n| `--verbose` | — | Print per-run details including response content. |\n\n### Comparing two remote builds\n\n```bash\n# Compare 1.110.0 against 1.115.0 (no local build needed):\nnpm run perf:chat -- --build 1.110.0 --baseline-build 1.115.0 --runs 5\n```\n\n### Comparing two local builds\n\nBoth `--build` and `--baseline-build` accept local paths to VS Code executables. This enables apples-to-apples comparisons between any two builds:\n\n```bash\n# Compare two dev builds (e.g. feature branch vs main):\nnpm run perf:chat -- \\\n  --build .build\u002Felectron\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --baseline-build \u002Fpath\u002Fto\u002Fother\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --runs 5\n\n# Compare two production builds:\nnpm run perf:chat -- \\\n  --build ..\u002FVSCode-darwin-arm64-feature\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --baseline-build ..\u002FVSCode-darwin-arm64-main\u002FCode\\ -\\ OSS.app\u002FContents\u002FMacOS\u002FCode\\ -\\ OSS \\\n  --runs 5\n```\n\nLocal path baselines are never cached (the build may change between runs). Version string baselines are cached for reuse.\n\n### Build modes and mismatch detection\n\nThe tool classifies builds into three modes based on the executable path:\n\n| Mode | Source | Characteristics |\n|---|---|---|\n| `dev` | `.build\u002Felectron\u002F` (local dev) | Unbundled sources, `VSCODE_DEV=1`, `NODE_ENV=development`. Higher memory and startup overhead. |\n| `production` | `..\u002FVSCode-\u003Cplatform>-\u003Carch>\u002F` (from `gulp vscode`) | Bundled JS, no dev flags. Matches release characteristics but uses local source. |\n| `release` | `.vscode-test\u002F` (downloaded via `@vscode\u002Ftest-electron`) | Official published build. |\n\nWhen test and baseline builds have different modes (e.g. dev vs release), the tool shows a warning and prompts for confirmation. Use `--force` or `--ci` to skip the prompt.\n\nUsing `--production-build` builds a local bundled package via `gulp vscode` for fair comparison against a release baseline. This eliminates dev-mode overhead while still testing your local changes.\n\n```bash\n# Production build vs release baseline (fair comparison):\nnpm run perf:chat -- --production-build --baseline-build 1.115.0 --runs 5\n```\n\n### Settings overrides\n\nUse `--setting`, `--test-setting`, and `--baseline-setting` to inject VS Code settings into the launched instance. This is useful for A\u002FB testing experimental features:\n\n```bash\n# Enable a feature for the test build only:\nnpm run perf:chat -- --test-setting chat.experimental.incrementalRendering.enabled=true --runs 3\n\n# Compare two builds with different settings:\nnpm run perf:chat -- \\\n  --baseline-build \"..\u002Fvscode2\u002F.build\u002Felectron\u002FCode - OSS.app\u002FContents\u002FMacOS\u002FCode - OSS\" \\\n  --baseline-setting chat.experimental.incrementalRendering.enabled=true \\\n  --test-setting chat.experimental.incrementalRendering.enabled=false \\\n  --runs 3\n\n# Set a value for both builds:\nnpm run perf:chat -- --setting chat.mcp.enabled=false --runs 3\n```\n\nPrecedence: `--test-setting` \u002F `--baseline-setting` override `--setting` for the same key. Values are auto-parsed: `true`\u002F`false` become booleans, numbers become numbers, everything else stays a string.\n\n### Resuming a run for more confidence\n\nWhen results exceed the threshold but aren't statistically significant, the tool prints a `--resume` hint. Use it to add more iterations to an existing run:\n\n```bash\n# Initial run with 3 iterations — may be inconclusive:\nnpm run perf:chat -- --scenario text-only --runs 3\n\n# Add 3 more runs to the same results file (both test + baseline):\nnpm run perf:chat -- --resume .chat-simulation-data\u002F2026-04-14T02-15-14\u002Fresults.json --runs 3\n\n# Keep adding until confidence is reached:\nnpm run perf:chat -- --resume .chat-simulation-data\u002F2026-04-14T02-15-14\u002Fresults.json --runs 5\n```\n\n`--resume` loads the previous `results.json` and its associated `baseline-*.json`, runs N more iterations for both builds, merges rawRuns, recomputes stats, and re-runs the comparison. The updated files are written back in-place. You can resume multiple times — samples accumulate.\n\n### Statistical significance\n\nRegression detection uses **Welch's t-test** to avoid false positives from noisy measurements. A metric is only flagged as `REGRESSION` when it both exceeds the threshold AND is statistically significant (p \u003C 0.05). Otherwise it's reported as `(likely noise — p=X, not significant)`.\n\nWith typical variance (cv ≈ 20%), you need:\n- **n ≥ 5** per build to detect a 35% regression at 95% confidence\n- **n ≥ 10** per build to detect a 20% regression reliably\n\nConfidence levels reported: `high` (p \u003C 0.01), `medium` (p \u003C 0.05), `low` (p \u003C 0.1), `none`.\n\n### Exit codes\n\n- `0` — all metrics within threshold, or exceeding threshold but not statistically significant\n- `1` — statistically significant regression detected, or all runs failed\n\n### Scenarios\n\nScenarios are defined in `scripts\u002Fchat-simulation\u002Fcommon\u002Fperf-scenarios.js` and registered via `registerPerfScenarios()`. There are three categories:\n\n- **Content-only** — plain streaming responses (e.g. `text-only`, `large-codeblock`, `rapid-stream`)\n- **Tool-call** — multi-turn scenarios with tool invocations (e.g. `tool-read-file`, `tool-edit-file`)\n- **Multi-turn user** — multi-turn conversations with user follow-ups, thinking blocks (e.g. `thinking-response`, `multi-turn-user`, `long-conversation`)\n\nRun `npm run perf:chat -- --help` to see the full list of registered scenario IDs.\n\n### Metrics collected\n\n- **Timing:** time to first token, time to complete, time to render complete (includes typewriter animation)\n- **Rendering:** layout count, layout duration (ms), style recalculation count, forced reflows, long tasks (>50ms), long animation frame count and duration\n- **Memory:** heap before\u002Fafter, heap delta post-GC (informational, noisy for single requests)\n- **Extension host:** heap before\u002Fafter\u002Fdelta via CDP inspector\n\n### Regression triggers vs informational metrics\n\nOnly these metrics trigger a regression failure (when they exceed the threshold with statistical significance):\n- `timeToFirstToken`, `timeToComplete` — user-perceived latency\n- `layoutDurationMs` — total layout time from the trace (the *real* layout cost)\n- `forcedReflowCount` — forced synchronous layouts are always bad\n- `longTaskCount` — main thread jank\n\nThese are reported but **informational only** (won't fail CI):\n- `longAnimationFrameCount` — number of long animation frames; noisy and compositor-driven, so tracked for signal but does not gate. Real jank is gated via `longTaskCount` \u002F `layoutDurationMs`.\n- `layoutCount` — number of layout ops; inflated by CSS animations (compositor-driven, cheap). A build can do *more but cheaper* layouts, so gate on `layoutDurationMs`, not this count.\n- `recalcStyleCount` — number of style recalcs; inflated by CSS animations (compositor-driven, cheap)\n- `timeToRenderComplete` — includes typewriter animation tail\n- Memory\u002Fheap metrics — too noisy for single-request benchmarks\n\n### Statistics\n\nResults use **IQR-based outlier removal** and **median** (not mean) to handle startup jitter. The **coefficient of variation (cv)** is reported — under 15% is stable, over 15% gets a ⚠ warning. Baseline comparison uses **Welch's t-test** on raw run values to determine statistical significance before flagging regressions. Use 5+ runs to get stable results.\n\n## Memory leak check\n\n**Script:** `scripts\u002Fchat-simulation\u002Ftest-chat-mem-leaks.js` **npm:** `npm run perf:chat-leak`\n\nLaunches one VS Code session, sends N messages sequentially, forces GC between each, and measures renderer heap and DOM node count. Uses **linear regression** on the samples to compute per-message growth rate, which is compared against a threshold.\n\n### Key flags\n\n| Flag | Default | Description |\n|---|---|---|\n| `--messages \u003Cn>` \u002F `-n` | `10` | Number of messages to send. More = more accurate slope. |\n| `--build \u003Cpath\\|ver>` \u002F `-b` | local dev | Build to test. |\n| `--threshold \u003CMB>` | `2` | Max per-message heap growth in MB. |\n| `--setting \u003Ck=v>` | — | Set a VS Code setting override (repeatable). |\n| `--verbose` | — | Print per-message heap\u002FDOM counts. |\n\n### What it measures\n\n- **Heap growth slope** (MB\u002Fmessage) — linear regression over forced-GC heap samples. A leak shows as sustained positive slope.\n- **DOM node growth** (nodes\u002Fmessage) — catches rendering leaks where elements aren't cleaned up. Healthy chat virtualizes old messages so node count plateaus.\n\n### Interpreting results\n\n- `0.3–1.0 MB\u002Fmsg` — normal (V8 internal overhead, string interning)\n- `>2.0 MB\u002Fmsg` — likely leak, investigate retained objects\n- DOM nodes stable after first message — normal (chat list virtualization working)\n- DOM nodes growing linearly — rendering leak, check disposable cleanup\n\n## CI runs & pinpointing regressions\n\nThe perf + leak checks run in CI via the **`.github\u002Fworkflows\u002Fchat-perf.yml`** workflow (a scheduled daily `workflow_dispatch`, plus manual dispatch). Each run benchmarks the current `main` as the **test** build against a fixed release **baseline** (from `config.jsonc`, e.g. `1.122.0`). Because the baseline is fixed, the **test** median for a metric across successive daily runs traces `main`'s trajectory over time — that's what lets you bisect *when* something regressed or went flaky.\n\n### Finding and reading historic runs\n\n```bash\n# List recent runs (most recent first) — note the run IDs and dates\ngh run list --workflow chat-perf.yml -R microsoft\u002Fvscode --limit 30 \\\n  --json databaseId,status,conclusion,createdAt,headBranch \\\n  --jq '.[] | [.databaseId, (.conclusion\u002F\u002F.status), .createdAt, .headBranch] | @tsv'\n\n# See a run's per-job results\ngh run view \u003Crun-id> -R microsoft\u002Fvscode --json jobs \\\n  --jq '.jobs[] | [.name, (.conclusion\u002F\u002F.status)] | @tsv'\n\n# Trigger a run manually against any ref\u002Fversion:\ngh workflow run chat-perf.yml -R microsoft\u002Fvscode --ref main \\\n  -f test_build=\u003Cbranch|sha|version> -f baseline_build=\u003Cversion>\n```\n\n### Artifacts per run (and retention — this matters for old runs)\n\n| Artifact | Contents | Retention |\n|---|---|---|\n| `chat-perf-summary` | Unified `ci-summary.md`: verdicts, per-metric medians ±stddev, **per-run raw tables** | 30 days |\n| `perf-results-\u003Cgroup>` | Everything below **plus traces, CPU profiles, heap snapshots** (large) | 30 days |\n| `leak-results` | Leak log + `chat-simulation-leak-results.json` | 30 days |\n| `perf-summary-\u003Cgroup>` | `results.json` (full per-run metrics incl. `rawRuns`) + `baseline-*.json` | **1 day** |\n\n```bash\n# List a run's artifacts + whether they've expired\ngh api repos\u002Fmicrosoft\u002Fvscode\u002Factions\u002Fruns\u002F\u003Crun-id>\u002Fartifacts \\\n  --jq '.artifacts[] | [.name, .expired] | @tsv'\n\n# Download the human-readable summary (best first stop; survives 30 days)\ngh run download \u003Crun-id> -R microsoft\u002Fvscode -n chat-perf-summary\n```\n\n### Pinpointing where a metric regressed \u002F went flaky\n\n1. **Bisect across dates.** Pull `chat-perf-summary\u002Fci-summary.md` from several runs spanning the window. Compare the **test** median (and ±stddev) for the suspect metric+scenario run-to-run — the day it jumps is when `main` changed. A metric that goes *bimodal* (e.g. a raw-run column showing two clusters like `~250 \u002F ~900`) is the flaky signature; a stable shift is a genuine regression.\n2. **Confirm it's real vs. measurement noise.** Check the per-run raw tables (in `ci-summary.md`) — high ±stddev \u002F bimodal values mean the *median* is being pulled around by a few outlier runs, not a uniform slowdown.\n3. **Deep-dive the cause.** Download `perf-results-\u003Cgroup>` (has the `trace.json` per run) for a slow run and inspect what dominates the slow window. Trick that found the `gc_stats` artifact: sum main-thread `RunTask` durations between two `code\u002Fchat\u002F*` marks (e.g. `willCollectInstructions` → `didCollectInstructions`) — if the window is ~0% busy, it's an async wait or a GC pause, not real work; then look at the largest `X`-phase events in that window (`MajorGC`, `Layout`, etc.).\n4. **Reproduce a suspect commit locally** to bisect precisely: `npm run perf:chat -- --build \u003Csha> --baseline-build \u003Cver> --runs 7` (or two commits directly via `--build \u003CshaA> --baseline-build \u003CshaB>`).\n\n> Tip: `perf-summary-*` (the machine-readable `results.json` with `rawRuns`) is deleted after **1 day**, so for older runs rely on `chat-perf-summary` (raw tables, 30 days) or extract `results.json` from `perf-results-*` (also 30 days).\n\n## Architecture\n\n```\nscripts\u002Fchat-simulation\u002F\n├── common\u002F\n│   ├── mock-llm-server.js    # Mock CAPI server matching @vscode\u002Fcopilot-api URL structure\n│   ├── perf-scenarios.js     # Built-in scenario definitions (content, tool-call, multi-turn)\n│   └── utils.js              # Shared: paths, env setup, stats, launch helpers\n├── config.jsonc              # Default config (baseline version, runs, thresholds)\n├── fixtures\u002F                 # TypeScript fixture files used by tool-call scenarios\n├── test-chat-perf-regression.js\n└── test-chat-mem-leaks.js\n```\n\n### Mock server\n\nThe mock LLM server (`common\u002Fmock-llm-server.js`) implements the full CAPI URL structure from `@vscode\u002Fcopilot-api`'s `DomainService`:\n\n- `GET \u002Fmodels` — returns model metadata\n- `POST \u002Fmodels\u002Fsession` — returns `AutoModeAPIResponse` with `available_models` and `session_token`\n- `POST \u002Fmodels\u002Fsession\u002Fintent` — model router\n- `POST \u002Fchat\u002Fcompletions` — SSE streaming response matching the scenario\n- Agent, session, telemetry, and token endpoints\n\nThe copilot extension connects to this server via `IS_SCENARIO_AUTOMATION=1` mode with `overrideCapiUrl` and `overrideProxyUrl` settings. The `vscode-api-tests` extension must be disabled (`--disable-extension=vscode.vscode-api-tests`) because it contributes a duplicate `copilot` vendor that blocks the real extension's language model provider registration.\n\n### Adding a scenario\n\n1. Add a new entry to the appropriate object (`CONTENT_SCENARIOS`, `TOOL_CALL_SCENARIOS`, or `MULTI_TURN_SCENARIOS`) in `common\u002Fperf-scenarios.js` using the `ScenarioBuilder` API from `common\u002Fmock-llm-server.js`\n2. The scenario is auto-registered by `registerPerfScenarios()` — no manual ID list to update\n3. Run: `npm run perf:chat -- --scenario your-new-scenario --runs 1 --no-baseline --verbose`\n\n## Related skills\n\n- **heap-snapshot-analysis** — When a perf regression or leak check identifies high memory growth, use the heap-snapshot-analysis skill to dig deeper. It can parse `.heapsnapshot` files, compare before\u002Fafter snapshots, group object deltas, and trace retainer paths to find what keeps disposed objects alive. The chat-perf leak check measures overall heap slope; heap-snapshot-analysis finds the specific objects responsible.\n- **auto-perf-optimize** — For launching VS Code, driving a scenario, and capturing heap snapshots or CPU profiles automatically before doing low-level analysis.\n",{"name":5,"description":7},"main","af285532414d1fa855b4360d86103c12253e6f2f",".github\u002Fskills\u002Fchat-perf",0,"https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide","https:\u002F\u002Fgithub.com\u002FDragonBallerZ","https:\u002F\u002Fwww.velloralabs.com","Building Dangerously Disruptive Technology. ","2026-10-23 14:19:51",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add DragonBallerZ\u002Fdragon-ide --skill chat-perf",{"tool":39,"command":40},"bunx","bunx skills add DragonBallerZ\u002Fdragon-ide --skill chat-perf",{"tool":42,"command":43},"pnpm","pnpm dlx skills add DragonBallerZ\u002Fdragon-ide --skill chat-perf","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fchat-perf\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M7DNGQPVJGHW226R8PS3R31F","update-skills","Create or update repository skills and instructions when major learnings are discovered during a session. Use when the user says \"learn!\", when a significant pattern or pitfall is identified, or when reusable domain knowledge should be captured for future sessions.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Fupdate-skills\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M7DNGP5XJWV3CG63DVS6NR8Y","update-pr","Update the pull request for the current session. Use when the user wants to push new changes to an existing PR.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Fupdate-pr\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M7DNGJEMNJZR24NBJ7MF5V8R","troubleshoot","Investigate unexpected behavior in the current Copilot agent session by analyzing its event log. Use when the user asks why something happened, why a request was slow, why a tool was or was not used, or why instructions\u002Fskills\u002Fagents did not load.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Ftroubleshoot\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7DNGFVCVMFN5J4ZKV27CDR1","sync","Sync the current session branch with its upstream branch, or publish the current session branch to a remote. Use when the user asks to sync a branch, pull latest changes, rebase onto upstream, push current branch, publish branch, or set upstream.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Fsync\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7DNGC9GE8EKQED3ZR3KAJ0K","sync-upstream","Update a stale session branch by rebasing onto the latest origin. Use when the upstream has moved significantly and the session needs to catch up, resolving conflicts by preserving upstream changes and adapting session work to fit.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Fsync-upstream\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M7DNG9NJNG1QPJB6C3GFDHVS","merge","Merge changes from the topic branch to the merge base branch. Use when the user wants to merge their session's work back to the base branch.","https:\u002F\u002Fgithub.com\u002FDragonBallerZ\u002Fdragon-ide\u002Fblob\u002Fmain\u002Fsrc\u002Fvs\u002Fsessions\u002Fskills\u002Fmerge\u002FSKILL.md",[],[]]