[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M698QH7JVVA09769ER9P2TKY":3,"skill-same:01M698QH7JVVA09769ER9P2TKY":74},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":22,"publishedAt":40,"linkOnly":41,"markdownContent":42,"markdownBody":43,"frontMatter":44,"refName":45,"commitSha":46,"skillDirectory":47,"forksCount":48,"repositoryHtmlUrl":49,"repositoryHomepageUrl":50,"ownerHtmlUrl":51,"ownerBlogUrl":50,"ownerDescription":52,"updatedAt":53,"install":54,"resources":67},"01M698QH7JVVA09769ER9P2TKY","openclaw-proof-validation",null,"Plan and collect OpenClaw Windows validation\u002Fproof: tests, rubber-duck review, UI evidence, MCP output, and gateway runtime proof.","openclaw\u002Fopenclaw-windows-node","openclaw","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F252820863?v=4","01M698P3FSAZ6B4R928ZF19KPT","01M698Q6SND6RHZDPP8ZM0NSP7",2198,"MIT","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fopenclaw-proof-validation\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[9,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39],"openclaw proof validation","validation","proof","duck","ui","ui evidence","plan","output","review","runtime proof","tests","collect","runtime","evidence","windows","gateway","MCP","2026-11-10 10:12:30",false,"---\nname: openclaw-proof-validation\ndescription: \"Plan and collect OpenClaw Windows validation\u002Fproof: tests, rubber-duck review, UI evidence, MCP output, and gateway runtime proof.\"\n---\n\n# OpenClaw Proof and Validation\n\nUse for changes that affect tray UX, Settings, onboarding, chat\u002Fcanvas, Command Center, Windows node capabilities, local MCP, gateway connection\u002Fpairing, permissions, diagnostics, or agent-facing instructions.\n\nIf the validation host is macOS, read [PARALLELS.md](PARALLELS.md) for the optional local Parallels Windows VM workflow.\n\n## Rules\n\n- Required automated\u002Ffocused tests are mandatory. Do not ask to skip them.\n- Prefer isolated tray data so proof does not mutate `%APPDATA%\\OpenClawTray`.\n- Computer-use is usually a batched closeout proof pass, not a continuous dev-loop tool. Mid-development use is fine when explicitly requested or needed to unblock work; ask first whether to run it or provide manual screenshot\u002Foutput steps.\n- Local MCP is part of every Windows node command contract: command discovery and invocation must work through `winnode` or raw MCP JSON-RPC.\n- Rubber-duck review is required before PR publication for non-trivial UI\u002FMCP\u002Fnode-command\u002Fsetup\u002Fpairing\u002Fsecurity\u002Fpermissions\u002Fdiagnostics work; it is also useful mid-development when extra design\u002Ftesting validation is requested.\n- Report blockers explicitly. Do not turn missing UI, MCP, gateway, camera, screen, or permission proof into success-shaped wording.\n\n## Required validation\n\nFirst follow [execution-target routing](..\u002F..\u002F..\u002Fdocs\u002FTEST_COVERAGE.md#choose-the-local-execution-target):\napp-only proof defaults to the isolated fixture Gateway; native-Gateway E2E\ndefaults to [named side-by-side patches](..\u002Fnative-gateway-from-source\u002FSKILL.md#default-multi-instance-native-gateway-testing)\non an authorized, capable Windows client host, locally or in future cloud lanes.\nUse separate VMs for OS\u002Fclean-image isolation or unavailable host capabilities,\nnot one VM per Gateway. Preserve required real MXC\u002FWSL proof and explicit\nauthority for existing shared Gateways.\n\nSelect the changed contract before running checks. Use the\n[read-only validation plan](..\u002F..\u002F..\u002Fdocs\u002FTEST_COVERAGE.md#explain-the-validation-plan)\nto distinguish hosted lanes, mandatory local code-change closeout, and custom\nproof candidates. A candidate is neither scheduled nor executed proof.\nOrdinary prose-only changes need documentation\u002Fskill checks, not native runtime\ntests. Shipped `winnode` command-reference changes retain CLI\u002Fcore checks.\n\n```powershell\n$env:OPENCLAW_REPO_ROOT = (Get-Location).Path\n.\\build.ps1\ndotnet test .\\tests\\OpenClaw.Shared.Tests\\OpenClaw.Shared.Tests.csproj --no-restore\ndotnet test .\\tests\\OpenClaw.Tray.Tests\\OpenClaw.Tray.Tests.csproj --no-restore\n```\n\nFresh worktrees may need a first run without `--no-restore`, or a project build first, so tests do not no-op before `bin\\` exists.\nCheck actual report\u002Fexecution counts, not only the exit code. Custom proof must\nnot contain skipped selected cases. Ordinary CI skips remain a separate policy,\nnot proof that the skipped behavior works.\n\nFor `winnode`, command descriptions, or new\u002Frenamed node commands, also run:\n\n```powershell\ndotnet test .\\tests\\OpenClaw.WinNode.Cli.Tests\\OpenClaw.WinNode.Cli.Tests.csproj --no-restore\n```\n\n## Proof checklist\n\n### Select the runtime and its authority\n\nOrdinary prose validation does not need an app. For supported deterministic app\nexploration, prefer the [fixture workflow](..\u002F..\u002F..\u002Fdocs\u002FGATEWAY_FIXTURE_TESTING.md#build-and-explore)\nwith an explicit current-build executable on an authorized Windows desktop.\nIt uses synthetic Gateway data, isolated profiles and a dedicated MCP port,\nwith owned-process cleanup. Sensitive node capabilities, automatic repair and\nvoice are disabled. It is not an OS sandbox, node-command proof, or real\nGateway pairing\u002Fprovider proof. An interactive `ready` result is not a passed\nscenario.\n\nFor live app proof outside that fixture, select the Gateway before launching.\nFor an isolated build that needs a real native Gateway, the default is the\n[task-owned build and named-patch workflow](..\u002Fnative-gateway-from-source\u002FSKILL.md#prepare-an-isolated-test-build):\nbuild Companion in its own worktree, prepare a uniquely named Gateway patch in\na task-owned work root, and launch with that patch and separate app data.\nThe launcher's shared `npm-latest` \u002F `packaged` defaults are not per-test\ninstances; `-Isolated` alone does not create a unique Gateway.\nThe [local launcher](..\u002F..\u002F..\u002Frun-app-local.ps1) has these distinct effects:\n\n| Option | Actual boundary |\n|---|---|\n| `-Isolated` \u002F `-DataDir` | Selects separate profile data, not an OS sandbox or permission to change machine runtimes. `-Isolated` uses a stable worktree\u002Fbranch directory, not a fresh profile on every run. |\n| Default Gateway (no explicit selection) | Builds\u002Fregisters an npm-latest development native Gateway, or reuses an existing patch selected by `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH` without rebuilding. |\n| `-GatewayChannel Latest` | Builds\u002Fregisters an npm-latest development native Gateway. Conflicts with a nonblank `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH`; clear that variable before selecting an explicit channel. |\n| `-GatewayChannel Packaged` | Builds\u002Fregisters a development Gateway from a published packaging release, not the installed Store runtime. Like `Latest`, conflicts with a nonblank `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH`. |\n| `-UseStoreGateway` | Skips development Gateway packaging and selects the installed Store Gateway. It does not make the launched app read-only. |\n| `-NoBuild` | Skips only the Companion build. Requires existing output with matching identity\u002Fconfiguration\u002FRID; Gateway preparation can still run. |\n| `-DryRun` | Returns before Gateway packaging and app launch, but after the optional Companion build and profile-directory creation. Not a read-only planner, even with `-NoBuild`. |\n\nLocal implementation authority does not authorize Gateway registration, installs,\nor shared-runtime changes. Use the\n[native Gateway skill](..\u002Fnative-gateway-from-source\u002FSKILL.md) only with the\nneeded authority. Do not execute the launcher to discover its parameters;\ninspect its PowerShell AST for read-only command checks. There is no\n`-AllowNonMain` switch.\n\nFor example, only when use of the installed Store Gateway and a live app launch\nare authorized:\n\n```powershell\n.\\run-app-local.ps1 -Isolated -UseStoreGateway\n```\n\nExpect the launcher to print the selected profile, Gateway, output path and\nprocess identity. Record the actual build\u002Fruntime identity; isolation alone is\nnot evidence that the binary matches the current source.\n\n### Collect the selected proof\n\nUse the [WinApp CLI guide](..\u002F..\u002F..\u002Fdocs\u002FWINAPP_CLI.md) proactively when choosing\nnative evidence: UIA state plus visible capture, stowed-exception diagnostics,\nDevTools inspection, bounded WinUI ETW, or owned Sandbox app runs. Discover\nthe installed version and command schema instead of assuming those tools are\nabsent. Preserve the runtime-selection rules above and the host's computer-use\npolicy; WinApp does not bypass consent or substitute for existing proof pools.\nRecord trace coverage and capture completeness, not just a successful command.\nFor clean artifact\u002Finstaller proof, consult the\n[sandbox-mcp versus WinApp decision table](..\u002F..\u002F..\u002Fdocs\u002FWINAPP_CLI.md#choose-winapp-sandbox-or-sandbox-mcp).\nPreserve missing-prerequisite findings, label explicitly provisioned runs, and\ninspect reports\u002Fscreenshots rather than equating a startup PASS with acceptance.\nThe two tools must not compete for the same guest's lifecycle.\n\n| Surface | Proof to collect |\n|---|---|\n| UI \u002F WinUI | On an authorized isolated desktop, use the runtime selected above, exercise the changed path with computer-use or developer-provided screenshots\u002Foutput, and include visible evidence or blocker. If the developer captures manually, provide exact steps and confirm screenshot\u002Fartifact links resolve after an authorized PR body update. |\n| Local MCP | Enable **Local MCP Server**, run `winnode --list-tools`, then invoke the changed command with `winnode --command \u003Cname> --params '\u003Cjson-object>'`. |\n| Raw MCP HTTP | For protocol\u002Fserver-shape changes, paste JSON-RPC `tools\u002Flist` and `tools\u002Fcall` responses from `http:\u002F\u002F127.0.0.1:8765\u002F`. |\n| Gateway path | When relevant and available, prove `openclaw nodes invoke --command \u003Cname> --params '\u003Cjson-object>'`; otherwise state the gateway blocker. |\n| Rubber-duck | Inspect the current host's advertised tools\u002Froles. When `task` advertises `rubber-duck`, use it for a bounded independent implementation\u002Fproof-plan review. Otherwise use an advertised independent review role and name the actual tool\u002Frole and limits; if none exists, report independent review blocked. Direct self-review is not independent review. Verify findings before changing code. |\n\nFor shared source-review mode, base, P0-P3 scope and fallback requirements, use\nthe [autoreview entrypoint](..\u002Fautoreview\u002FSKILL.md). Host tool availability does\nnot establish model access or a completed review. Keep source review separate\nfrom live UI, MCP and Gateway proof.\n\n## When local MCP proof is refused\n\nConnection refused on `http:\u002F\u002F127.0.0.1:8765\u002F` means the tray MCP server is not running. Do not reinstall `winnode`. Write PascalCase `settings.json` with `EnableMcpServer` true and `EnableNodeMode` false, launch the tray with `OPENCLAW_TRAY_DATA_DIR` and `OPENCLAW_SKIP_UPDATE_CHECK=1`, and pass that same data directory to `winnode`. `StartupSetupState.RequiresSetup` stays false when MCP is already enabled, so the first-run wizard does not block. An isolated data directory uses its own mutex. Wait until `mcp-token.txt` exists and the port accepts a connection.\n\n`browser.proxy` stays off `tools\u002Flist` until the app has a gateway client, a non-empty shared gateway token, and a verified browser endpoint. `Unknown tool: browser.proxy` is that rule. A refused port is not successful discovery. The review can accept a concrete gateway-availability blocker. It does not accept the refused port as the `browser.proxy` call.\n\nWhen the review asks for the production HTTP client, drive that client against a real socket. `HttpClient.Timeout` does not cancel a body read that starts after `HttpCompletionOption.ResponseHeadersRead`.\n\nFor isolated tray runs, copy the data directory printed by `run-app-local.ps1 -Isolated` and set it before MCP proof commands:\n\n```powershell\n$env:OPENCLAW_TRAY_DATA_DIR = '\u003Cisolated-data-dir-from-run-app-local>'\n```\n\nRaw token lookup:\n\n```powershell\n$tokenPath = if ($env:OPENCLAW_TRAY_DATA_DIR) {\n    Join-Path $env:OPENCLAW_TRAY_DATA_DIR 'mcp-token.txt'\n} else {\n    Join-Path $env:APPDATA 'OpenClawTray\\mcp-token.txt'\n}\n$token = Get-Content $tokenPath -Raw\n```\n\n## New Windows node command checklist\n\n1. Register the command in the same `INodeCapability` path used by the gateway node.\n2. Add\u002Fupdate `McpToolBridge.CommandDescriptions`.\n3. Update `.agents\u002Fskills\u002Fwinnode\u002FSKILL.md` with input shape, output shape, side effects, permissions, and examples.\n4. Add\u002Fupdate capability, MCP bridge, `winnode`, and UI\u002Fgateway tests as applicable.\n5. Prove discovery and invocation with `winnode` or raw MCP JSON-RPC.\n6. Prove the gateway path when the behavior is gateway-mediated and a gateway is available.\n\n## PR proof package\n\nThis section owns evidence requirements and their limits. Use\n[Windows PR preparation](..\u002Fwindows-pr-prepare\u002FSKILL.md) to assemble that evidence\ninto the current PR template and report readiness without assuming publication\nauthority.\n\nBefore publishing or updating a PR, collect:\n\n1. `## Validation` with exact commands and pass\u002Ffail counts.\n2. `## Real behavior proof` with current-head after-change evidence that directly shows the changed behavior: copied live output, screenshot\u002Fvideo, developer-provided screenshot, copied UI diagnostics, `winnode`, raw MCP JSON-RPC, gateway invoke output, redacted runtime log, or linked artifact. For UI changes, prefer screenshots\u002Fvideo of the active changed state, not only adjacent or empty UI.\n3. PR body proof must stay in sync with current behavior: remove stale screenshots\u002Fclaims after design changes and verify embedded image\u002Fartifact links resolve.\n4. Rubber-duck notes for non-trivial UI\u002FMCP-sensitive work, or why skipped.\n5. `Not verified \u002F blocked` notes for focused proof or unavailable dependencies.\n\nFor inventory-bound filtered tests, use\n[local execution evidence validation](..\u002F..\u002F..\u002Fdocs\u002FPROOF_POOLS.md#validate-local-execution-evidence).\nThe summary checks source\u002Freport\u002Fartifact correspondence; it does not replace\nnative freeze\u002Fbinary verification, visible UI inspection, or live Gateway proof,\nand it never marks a whole pool complete.\n","# OpenClaw Proof and Validation\n\nUse for changes that affect tray UX, Settings, onboarding, chat\u002Fcanvas, Command Center, Windows node capabilities, local MCP, gateway connection\u002Fpairing, permissions, diagnostics, or agent-facing instructions.\n\nIf the validation host is macOS, read [PARALLELS.md](PARALLELS.md) for the optional local Parallels Windows VM workflow.\n\n## Rules\n\n- Required automated\u002Ffocused tests are mandatory. Do not ask to skip them.\n- Prefer isolated tray data so proof does not mutate `%APPDATA%\\OpenClawTray`.\n- Computer-use is usually a batched closeout proof pass, not a continuous dev-loop tool. Mid-development use is fine when explicitly requested or needed to unblock work; ask first whether to run it or provide manual screenshot\u002Foutput steps.\n- Local MCP is part of every Windows node command contract: command discovery and invocation must work through `winnode` or raw MCP JSON-RPC.\n- Rubber-duck review is required before PR publication for non-trivial UI\u002FMCP\u002Fnode-command\u002Fsetup\u002Fpairing\u002Fsecurity\u002Fpermissions\u002Fdiagnostics work; it is also useful mid-development when extra design\u002Ftesting validation is requested.\n- Report blockers explicitly. Do not turn missing UI, MCP, gateway, camera, screen, or permission proof into success-shaped wording.\n\n## Required validation\n\nFirst follow [execution-target routing](..\u002F..\u002F..\u002Fdocs\u002FTEST_COVERAGE.md#choose-the-local-execution-target):\napp-only proof defaults to the isolated fixture Gateway; native-Gateway E2E\ndefaults to [named side-by-side patches](..\u002Fnative-gateway-from-source\u002FSKILL.md#default-multi-instance-native-gateway-testing)\non an authorized, capable Windows client host, locally or in future cloud lanes.\nUse separate VMs for OS\u002Fclean-image isolation or unavailable host capabilities,\nnot one VM per Gateway. Preserve required real MXC\u002FWSL proof and explicit\nauthority for existing shared Gateways.\n\nSelect the changed contract before running checks. Use the\n[read-only validation plan](..\u002F..\u002F..\u002Fdocs\u002FTEST_COVERAGE.md#explain-the-validation-plan)\nto distinguish hosted lanes, mandatory local code-change closeout, and custom\nproof candidates. A candidate is neither scheduled nor executed proof.\nOrdinary prose-only changes need documentation\u002Fskill checks, not native runtime\ntests. Shipped `winnode` command-reference changes retain CLI\u002Fcore checks.\n\n```powershell\n$env:OPENCLAW_REPO_ROOT = (Get-Location).Path\n.\\build.ps1\ndotnet test .\\tests\\OpenClaw.Shared.Tests\\OpenClaw.Shared.Tests.csproj --no-restore\ndotnet test .\\tests\\OpenClaw.Tray.Tests\\OpenClaw.Tray.Tests.csproj --no-restore\n```\n\nFresh worktrees may need a first run without `--no-restore`, or a project build first, so tests do not no-op before `bin\\` exists.\nCheck actual report\u002Fexecution counts, not only the exit code. Custom proof must\nnot contain skipped selected cases. Ordinary CI skips remain a separate policy,\nnot proof that the skipped behavior works.\n\nFor `winnode`, command descriptions, or new\u002Frenamed node commands, also run:\n\n```powershell\ndotnet test .\\tests\\OpenClaw.WinNode.Cli.Tests\\OpenClaw.WinNode.Cli.Tests.csproj --no-restore\n```\n\n## Proof checklist\n\n### Select the runtime and its authority\n\nOrdinary prose validation does not need an app. For supported deterministic app\nexploration, prefer the [fixture workflow](..\u002F..\u002F..\u002Fdocs\u002FGATEWAY_FIXTURE_TESTING.md#build-and-explore)\nwith an explicit current-build executable on an authorized Windows desktop.\nIt uses synthetic Gateway data, isolated profiles and a dedicated MCP port,\nwith owned-process cleanup. Sensitive node capabilities, automatic repair and\nvoice are disabled. It is not an OS sandbox, node-command proof, or real\nGateway pairing\u002Fprovider proof. An interactive `ready` result is not a passed\nscenario.\n\nFor live app proof outside that fixture, select the Gateway before launching.\nFor an isolated build that needs a real native Gateway, the default is the\n[task-owned build and named-patch workflow](..\u002Fnative-gateway-from-source\u002FSKILL.md#prepare-an-isolated-test-build):\nbuild Companion in its own worktree, prepare a uniquely named Gateway patch in\na task-owned work root, and launch with that patch and separate app data.\nThe launcher's shared `npm-latest` \u002F `packaged` defaults are not per-test\ninstances; `-Isolated` alone does not create a unique Gateway.\nThe [local launcher](..\u002F..\u002F..\u002Frun-app-local.ps1) has these distinct effects:\n\n| Option | Actual boundary |\n|---|---|\n| `-Isolated` \u002F `-DataDir` | Selects separate profile data, not an OS sandbox or permission to change machine runtimes. `-Isolated` uses a stable worktree\u002Fbranch directory, not a fresh profile on every run. |\n| Default Gateway (no explicit selection) | Builds\u002Fregisters an npm-latest development native Gateway, or reuses an existing patch selected by `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH` without rebuilding. |\n| `-GatewayChannel Latest` | Builds\u002Fregisters an npm-latest development native Gateway. Conflicts with a nonblank `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH`; clear that variable before selecting an explicit channel. |\n| `-GatewayChannel Packaged` | Builds\u002Fregisters a development Gateway from a published packaging release, not the installed Store runtime. Like `Latest`, conflicts with a nonblank `OPENCLAW_NATIVE_GATEWAY_DEV_PATCH`. |\n| `-UseStoreGateway` | Skips development Gateway packaging and selects the installed Store Gateway. It does not make the launched app read-only. |\n| `-NoBuild` | Skips only the Companion build. Requires existing output with matching identity\u002Fconfiguration\u002FRID; Gateway preparation can still run. |\n| `-DryRun` | Returns before Gateway packaging and app launch, but after the optional Companion build and profile-directory creation. Not a read-only planner, even with `-NoBuild`. |\n\nLocal implementation authority does not authorize Gateway registration, installs,\nor shared-runtime changes. Use the\n[native Gateway skill](..\u002Fnative-gateway-from-source\u002FSKILL.md) only with the\nneeded authority. Do not execute the launcher to discover its parameters;\ninspect its PowerShell AST for read-only command checks. There is no\n`-AllowNonMain` switch.\n\nFor example, only when use of the installed Store Gateway and a live app launch\nare authorized:\n\n```powershell\n.\\run-app-local.ps1 -Isolated -UseStoreGateway\n```\n\nExpect the launcher to print the selected profile, Gateway, output path and\nprocess identity. Record the actual build\u002Fruntime identity; isolation alone is\nnot evidence that the binary matches the current source.\n\n### Collect the selected proof\n\nUse the [WinApp CLI guide](..\u002F..\u002F..\u002Fdocs\u002FWINAPP_CLI.md) proactively when choosing\nnative evidence: UIA state plus visible capture, stowed-exception diagnostics,\nDevTools inspection, bounded WinUI ETW, or owned Sandbox app runs. Discover\nthe installed version and command schema instead of assuming those tools are\nabsent. Preserve the runtime-selection rules above and the host's computer-use\npolicy; WinApp does not bypass consent or substitute for existing proof pools.\nRecord trace coverage and capture completeness, not just a successful command.\nFor clean artifact\u002Finstaller proof, consult the\n[sandbox-mcp versus WinApp decision table](..\u002F..\u002F..\u002Fdocs\u002FWINAPP_CLI.md#choose-winapp-sandbox-or-sandbox-mcp).\nPreserve missing-prerequisite findings, label explicitly provisioned runs, and\ninspect reports\u002Fscreenshots rather than equating a startup PASS with acceptance.\nThe two tools must not compete for the same guest's lifecycle.\n\n| Surface | Proof to collect |\n|---|---|\n| UI \u002F WinUI | On an authorized isolated desktop, use the runtime selected above, exercise the changed path with computer-use or developer-provided screenshots\u002Foutput, and include visible evidence or blocker. If the developer captures manually, provide exact steps and confirm screenshot\u002Fartifact links resolve after an authorized PR body update. |\n| Local MCP | Enable **Local MCP Server**, run `winnode --list-tools`, then invoke the changed command with `winnode --command \u003Cname> --params '\u003Cjson-object>'`. |\n| Raw MCP HTTP | For protocol\u002Fserver-shape changes, paste JSON-RPC `tools\u002Flist` and `tools\u002Fcall` responses from `http:\u002F\u002F127.0.0.1:8765\u002F`. |\n| Gateway path | When relevant and available, prove `openclaw nodes invoke --command \u003Cname> --params '\u003Cjson-object>'`; otherwise state the gateway blocker. |\n| Rubber-duck | Inspect the current host's advertised tools\u002Froles. When `task` advertises `rubber-duck`, use it for a bounded independent implementation\u002Fproof-plan review. Otherwise use an advertised independent review role and name the actual tool\u002Frole and limits; if none exists, report independent review blocked. Direct self-review is not independent review. Verify findings before changing code. |\n\nFor shared source-review mode, base, P0-P3 scope and fallback requirements, use\nthe [autoreview entrypoint](..\u002Fautoreview\u002FSKILL.md). Host tool availability does\nnot establish model access or a completed review. Keep source review separate\nfrom live UI, MCP and Gateway proof.\n\n## When local MCP proof is refused\n\nConnection refused on `http:\u002F\u002F127.0.0.1:8765\u002F` means the tray MCP server is not running. Do not reinstall `winnode`. Write PascalCase `settings.json` with `EnableMcpServer` true and `EnableNodeMode` false, launch the tray with `OPENCLAW_TRAY_DATA_DIR` and `OPENCLAW_SKIP_UPDATE_CHECK=1`, and pass that same data directory to `winnode`. `StartupSetupState.RequiresSetup` stays false when MCP is already enabled, so the first-run wizard does not block. An isolated data directory uses its own mutex. Wait until `mcp-token.txt` exists and the port accepts a connection.\n\n`browser.proxy` stays off `tools\u002Flist` until the app has a gateway client, a non-empty shared gateway token, and a verified browser endpoint. `Unknown tool: browser.proxy` is that rule. A refused port is not successful discovery. The review can accept a concrete gateway-availability blocker. It does not accept the refused port as the `browser.proxy` call.\n\nWhen the review asks for the production HTTP client, drive that client against a real socket. `HttpClient.Timeout` does not cancel a body read that starts after `HttpCompletionOption.ResponseHeadersRead`.\n\nFor isolated tray runs, copy the data directory printed by `run-app-local.ps1 -Isolated` and set it before MCP proof commands:\n\n```powershell\n$env:OPENCLAW_TRAY_DATA_DIR = '\u003Cisolated-data-dir-from-run-app-local>'\n```\n\nRaw token lookup:\n\n```powershell\n$tokenPath = if ($env:OPENCLAW_TRAY_DATA_DIR) {\n    Join-Path $env:OPENCLAW_TRAY_DATA_DIR 'mcp-token.txt'\n} else {\n    Join-Path $env:APPDATA 'OpenClawTray\\mcp-token.txt'\n}\n$token = Get-Content $tokenPath -Raw\n```\n\n## New Windows node command checklist\n\n1. Register the command in the same `INodeCapability` path used by the gateway node.\n2. Add\u002Fupdate `McpToolBridge.CommandDescriptions`.\n3. Update `.agents\u002Fskills\u002Fwinnode\u002FSKILL.md` with input shape, output shape, side effects, permissions, and examples.\n4. Add\u002Fupdate capability, MCP bridge, `winnode`, and UI\u002Fgateway tests as applicable.\n5. Prove discovery and invocation with `winnode` or raw MCP JSON-RPC.\n6. Prove the gateway path when the behavior is gateway-mediated and a gateway is available.\n\n## PR proof package\n\nThis section owns evidence requirements and their limits. Use\n[Windows PR preparation](..\u002Fwindows-pr-prepare\u002FSKILL.md) to assemble that evidence\ninto the current PR template and report readiness without assuming publication\nauthority.\n\nBefore publishing or updating a PR, collect:\n\n1. `## Validation` with exact commands and pass\u002Ffail counts.\n2. `## Real behavior proof` with current-head after-change evidence that directly shows the changed behavior: copied live output, screenshot\u002Fvideo, developer-provided screenshot, copied UI diagnostics, `winnode`, raw MCP JSON-RPC, gateway invoke output, redacted runtime log, or linked artifact. For UI changes, prefer screenshots\u002Fvideo of the active changed state, not only adjacent or empty UI.\n3. PR body proof must stay in sync with current behavior: remove stale screenshots\u002Fclaims after design changes and verify embedded image\u002Fartifact links resolve.\n4. Rubber-duck notes for non-trivial UI\u002FMCP-sensitive work, or why skipped.\n5. `Not verified \u002F blocked` notes for focused proof or unavailable dependencies.\n\nFor inventory-bound filtered tests, use\n[local execution evidence validation](..\u002F..\u002F..\u002Fdocs\u002FPROOF_POOLS.md#validate-local-execution-evidence).\nThe summary checks source\u002Freport\u002Fartifact correspondence; it does not replace\nnative freeze\u002Fbinary verification, visible UI inspection, or live Gateway proof,\nand it never marks a whole pool complete.\n",{"name":5,"description":7},"main","a6ff6a164b20835bd5d77200c9134578dd7ce0c6",".agents\u002Fskills\u002Fopenclaw-proof-validation",349,"https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node","https:\u002F\u002Fopenclaw.ai","https:\u002F\u002Fgithub.com\u002Fopenclaw","Your personal, open source AI assistant.","2026-11-14 18:13:49",{"source":8,"skillName":5,"commands":55,"prompt":65,"zipAvailable":66},[56,59,62],{"tool":57,"command":58},"npx","npx skills add openclaw\u002Fopenclaw-windows-node --skill openclaw-proof-validation",{"tool":60,"command":61},"bunx","bunx skills add openclaw\u002Fopenclaw-windows-node --skill openclaw-proof-validation",{"tool":63,"command":64},"pnpm","pnpm dlx skills add openclaw\u002Fopenclaw-windows-node --skill openclaw-proof-validation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fopenclaw-proof-validation\u002FSKILL.md",true,[68],{"path":69,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":72,"sourceHtmlUrl":73},"PARALLELS.md","REFERENCE","参考文档",3471,"https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fopenclaw-proof-validation\u002FPARALLELS.md",[75,83,108,125,143,159],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":82},"01M7416EHCYH7YCJ9XPSQ2844V","network-handoff","Use before generating, changing, or reviewing networking, IPC, or cross-process handoffs, including requests, retries, deadlines, reconnect, process launch, configuration apply, cancellation, and cleanup. Establish confirmation, identity, authority, reconciliation, and producer compatibility before implementation.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fnetwork-handoff\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":40},"01M6KZCXK5PCYWWW5FCVDYPAXK","retrigger-flaky-ci","Triage failed CI on a pull request or the active branch, decide whether the failures come from its changes, and rerun only failures that are unrelated. Use when asked to retrigger, rerun, or kick CI on a PR with failing checks.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fretrigger-flaky-ci\u002FSKILL.md",[],[90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107],"ci","retrigger flaky ci","when","failing","rerun","use","pull request","triage","only","decide","request","active","pull","with","branch","unrelated","changes","pr",{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":40},"01M698QMQN5TQES4R7ZPTYAZF5","technical-documentation","Write, review, or restructure Windows node documentation and agent instructions using current source, task-focused explanations, and explicit fact preservation.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Ftechnical-documentation\u002FSKILL.md",[],[115,116,117,118,119,37,120,31,121,122,123,124],"technical documentation","technical","documentation","instructions","agent instructions","task","write","node","preservation","source",{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":131,"publishedAt":40},"01M698QKPXJ8ZG34MYHPPCAT4T","pr-simplify","Simplify your own Windows-node change to its smallest complete fix while preserving distinct regressions, platform contracts, and required proof. Use after implementation before PR preparation, after large repairs, or when asked to shrink a change or remove slop.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fpr-simplify\u002FSKILL.md",[],[132,107,133,122,134,135,25,136,92,95,137,138,139,140,141,37,142],"simplify","pr simplify","preparation","implementation","contracts","fix","complete","remove","slop","platform","change",{"skillNo":144,"name":145,"nameZh":6,"description":146,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":147,"categories":148,"keywords":152,"publishedAt":40},"01M698QJEA384YPM8R9CPZ1YFD","openclaw-testing","Choose proportional OpenClaw Windows node tests, builds, packaging checks, and runtime proof, then diagnose failures without weakening required gates.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fopenclaw-testing\u002FSKILL.md",[149],{"categoryNo":150,"categoryName":151,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000008","测试与质量",[153,9,154,37,25,122,155,32,33,156,157,35,158],"testing","openclaw testing","gates","choose","packaging","diagnose",{"skillNo":160,"name":161,"nameZh":6,"description":162,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":163,"categories":164,"keywords":170,"publishedAt":40},"01M698QFXW9PDK8YZ467MBW5P1","openclaw-debugging","Debug OpenClaw Windows Companion, setup, Gateway connection, node capabilities, packaged permissions, updates, installers, and system.run by selecting the failing boundary and proof path before changing code.","https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw-windows-node\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fopenclaw-debugging\u002FSKILL.md",[165],{"categoryNo":166,"categoryName":167,"taxonomyType":168,"taxonomyTypeDesc":169},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[9,171,172,173,174,25,175,176,177,178,37,38,179,122,180,181,182,183,184,93],"debugging","openclaw debugging","path","setup","connection","boundary","debug","system","companion","run","permissions","capabilities","code","updates"]