[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78WKAYG1FH1WBRN2HMSJS6K":3,"skill-same:01M78WKAYG1FH1WBRN2HMSJS6K":71},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":33,"install":34,"resources":46},"01M78WKAYG1FH1WBRN2HMSJS6K","statistics-and-extension",null,"Use Earth2Studio statistics, metrics, perturbations, and component protocols to build and validate safe custom inference extensions.","VectorSpaceLab\u002FAREX-Skill","VectorSpaceLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F180637102?v=4","01M7538C5WPSJGWF7KHR4EFZ9F","01M7538C6R2GFP53NPXTCZNBRT",351,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: statistics-and-extension\ndescription: \"Use Earth2Studio statistics, metrics, perturbations, and component\n  protocols to build and validate safe custom inference extensions.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# Statistics and extension\n\nUse this sub-skill when a request needs a reduction or verification metric, a\ncustom prognostic\u002Fdiagnostic\u002Fdata source\u002Fassimilation\u002Fperturbation component, or\na small workflow that composes those contracts. It targets Earth2Studio 0.18.0a0\npublic APIs and Python 3.11–3.14. It does not enumerate every model, data source,\nbackend, or optional extra.\n\n## Route the request\n\n- **Reduction:** one input, named `reduction_dimensions`; use `Statistic`.\n- **Verification:** forecast `x` plus truth `y`; use `Metric` and validate the\n  non-ensemble axes before arithmetic.\n- **Derived physical field:** use a diagnostic, not a statistic.\n- **Time stepping:** implement a prognostic and its initial-condition-first\n  iterator.\n- **Data access:** implement the matching array or DataFrame source protocol;\n  add a lexicon only when source keys need translation.\n- **Observation update:** implement the DA protocol with explicit schemas.\n- **Ensemble initial-state change:** implement a `Perturbation`; it receives and\n  returns a tensor\u002Fcoordinate pair.\n\nRead the focused contracts before editing:\n\n- [Statistics and metrics](references\u002Fstatistics-reference.md)\n- [Component extension contracts](references\u002Fextension-reference.md)\n- [Testing and validation](references\u002Ftesting-reference.md)\n- [Failure recovery](references\u002Ftroubleshooting.md)\n- [Offline contract smoke check](scripts\u002Fcontract_smoke.py)\n\n## Non-negotiable coordinate contract\n\nTreat a `CoordSystem` as an ordered mapping whose values describe tensor axes in\nexactly the same order. `batch` is the first dynamic model axis when using\n`batch_coords()`\u002F`batch_func()`. Common public model axes are `time`,\n`lead_time`, `variable`, `lat`, and `lon`; a prognostic has `lead_time`, while a\nsingle-step diagnostic normally does not. Do not silently reorder axes to make a\nbroadcast work.\n\nFor every custom component:\n\n1. Define `input_coords()` with fresh arrays and explicit Earth2Studio variable\n   names, physical units, and latitude\u002Flongitude conventions.\n2. Validate required dimension positions with `handshake_dim` and required values\n   with `handshake_coords` (or a deliberate equivalent for tabular schemas).\n3. Build `output_coords` from a copy, removing only reduced dimensions or changing\n   only the documented output fields. Every returned tensor axis needs a matching\n   coordinate variable of the same length.\n4. Keep data in physical units at component boundaries. Normalize only inside a\n   component and undo that normalization before returning.\n5. Test both a valid tiny tensor and a malformed coordinate mapping. Coordinate\n   shape equality alone is not enough when semantic axis order differs.\n\n## Minimal implementation loop\n\n1. Write the input\u002Foutput contract, axis order, variable vocabulary, units,\n   optional dependencies, device behavior, statefulness, and whether network or\n   checkpoints are required. Keep network\u002Fdownload actions out of smoke checks.\n2. Select the smallest protocol. A statistic only removes dimensions; a\n   diagnostic may add or transform variables; a prognostic must advance\n   `lead_time`; a source must return xarray or DataFrame data with the required\n   metadata; DA may return multiple state\u002Fobservation outputs.\n3. Implement the protocol and coordinate validation before the numerical method.\n   Use `@batch_coords()` and `@batch_func()` for ordinary tensor model patterns,\n   and use inference mode for inference-only perturbations\u002Fforwards.\n4. Compose through the verified workflow signatures. Use\n   `run.deterministic(time, nsteps, prognostic, data, io, output_coords=..., device=..., verbose=..., checkpoint=...)`,\n   `run.diagnostic(..., diagnostic, data, io, ...)`, or\n   `run.ensemble(..., nensemble, ..., perturbation, batch_size=..., ...)`.\n5. Validate output coordinates and tensor shapes at every boundary. For metrics,\n   compare forecast and truth axes by name and coordinate values before calling a\n   built-in reduction; an accidental same-shape subtraction can still be wrong.\n6. Run the offline smoke check, focused unit tests, then applicable native tests or\n   examples. Report skipped optional-backend\u002Fnetwork checks rather than claiming\n   support for them.\n\n## Common statistical recipes\n\n```python\nimport torch\n\nfrom earth2studio.statistics import lat_weight, mean, rmse\n\nweights = lat_weight(torch.as_tensor(latitudes))\narea_mean = mean([\"lat\"], weights=weights)\nscore = rmse([\"lat\", \"lon\"], weights=grid_weights,\n             ensemble_dimension=\"ensemble\")\n```\n\n`mean`, `variance`, and `std` can maintain state with `batch_update=True`.\nMetrics such as RMSE\u002FMAE can reduce an ensemble mean when\n`ensemble_dimension` is supplied. CRPS, rank histograms, and energy score expect\nan ensemble axis in forecast `x` and no such axis in truth `y`. Use the exact\nweight dimensionality and output-coordinate rules in the statistics reference.\n\n## Safe validation commands\n\nFrom any directory, run the bundled checker by its installed skill path:\n\n```bash\nuv run python \u003Cskill-root>\u002Fscripts\u002Fcontract_smoke.py --help\nuv run python \u003Cskill-root>\u002Fscripts\u002Fcontract_smoke.py --case all\n```\n\nWhen developing a custom component in your own project, run that project's\nfocused tests in its managed environment after the bundled contract smoke. A\nmissing optional Earth2Studio group is a limitation to report, not a reason to\ninstall an unbounded dependency set. See the testing reference for the local\nfixture expectations; the original repository's native test paths are review\nevidence, not runtime dependencies.\n\n## Scope boundaries and handoff\n\nThis sub-skill covers component implementation, statistics\u002Fmetric axes and\nweights, lexicon\u002Ftime-tolerance conventions, focused tests, and safe local\nvalidation. It intentionally excludes release\u002FCI operations, package-wide\nmaintainer workflows, serving-client construction, exhaustive catalogues, and\nremote data\u002Fmodel downloads. Preserve those limits in the downstream handoff.\nRecord native candidates and the two difficult synthetic cases: a diagnostic\nwhose returned coordinate mapping omits its output `variable`, and an ensemble\nmetric whose truth axes are reordered but retain a broadcastable shape.\n","# Statistics and extension\n\nUse this sub-skill when a request needs a reduction or verification metric, a\ncustom prognostic\u002Fdiagnostic\u002Fdata source\u002Fassimilation\u002Fperturbation component, or\na small workflow that composes those contracts. It targets Earth2Studio 0.18.0a0\npublic APIs and Python 3.11–3.14. It does not enumerate every model, data source,\nbackend, or optional extra.\n\n## Route the request\n\n- **Reduction:** one input, named `reduction_dimensions`; use `Statistic`.\n- **Verification:** forecast `x` plus truth `y`; use `Metric` and validate the\n  non-ensemble axes before arithmetic.\n- **Derived physical field:** use a diagnostic, not a statistic.\n- **Time stepping:** implement a prognostic and its initial-condition-first\n  iterator.\n- **Data access:** implement the matching array or DataFrame source protocol;\n  add a lexicon only when source keys need translation.\n- **Observation update:** implement the DA protocol with explicit schemas.\n- **Ensemble initial-state change:** implement a `Perturbation`; it receives and\n  returns a tensor\u002Fcoordinate pair.\n\nRead the focused contracts before editing:\n\n- [Statistics and metrics](references\u002Fstatistics-reference.md)\n- [Component extension contracts](references\u002Fextension-reference.md)\n- [Testing and validation](references\u002Ftesting-reference.md)\n- [Failure recovery](references\u002Ftroubleshooting.md)\n- [Offline contract smoke check](scripts\u002Fcontract_smoke.py)\n\n## Non-negotiable coordinate contract\n\nTreat a `CoordSystem` as an ordered mapping whose values describe tensor axes in\nexactly the same order. `batch` is the first dynamic model axis when using\n`batch_coords()`\u002F`batch_func()`. Common public model axes are `time`,\n`lead_time`, `variable`, `lat`, and `lon`; a prognostic has `lead_time`, while a\nsingle-step diagnostic normally does not. Do not silently reorder axes to make a\nbroadcast work.\n\nFor every custom component:\n\n1. Define `input_coords()` with fresh arrays and explicit Earth2Studio variable\n   names, physical units, and latitude\u002Flongitude conventions.\n2. Validate required dimension positions with `handshake_dim` and required values\n   with `handshake_coords` (or a deliberate equivalent for tabular schemas).\n3. Build `output_coords` from a copy, removing only reduced dimensions or changing\n   only the documented output fields. Every returned tensor axis needs a matching\n   coordinate variable of the same length.\n4. Keep data in physical units at component boundaries. Normalize only inside a\n   component and undo that normalization before returning.\n5. Test both a valid tiny tensor and a malformed coordinate mapping. Coordinate\n   shape equality alone is not enough when semantic axis order differs.\n\n## Minimal implementation loop\n\n1. Write the input\u002Foutput contract, axis order, variable vocabulary, units,\n   optional dependencies, device behavior, statefulness, and whether network or\n   checkpoints are required. Keep network\u002Fdownload actions out of smoke checks.\n2. Select the smallest protocol. A statistic only removes dimensions; a\n   diagnostic may add or transform variables; a prognostic must advance\n   `lead_time`; a source must return xarray or DataFrame data with the required\n   metadata; DA may return multiple state\u002Fobservation outputs.\n3. Implement the protocol and coordinate validation before the numerical method.\n   Use `@batch_coords()` and `@batch_func()` for ordinary tensor model patterns,\n   and use inference mode for inference-only perturbations\u002Fforwards.\n4. Compose through the verified workflow signatures. Use\n   `run.deterministic(time, nsteps, prognostic, data, io, output_coords=..., device=..., verbose=..., checkpoint=...)`,\n   `run.diagnostic(..., diagnostic, data, io, ...)`, or\n   `run.ensemble(..., nensemble, ..., perturbation, batch_size=..., ...)`.\n5. Validate output coordinates and tensor shapes at every boundary. For metrics,\n   compare forecast and truth axes by name and coordinate values before calling a\n   built-in reduction; an accidental same-shape subtraction can still be wrong.\n6. Run the offline smoke check, focused unit tests, then applicable native tests or\n   examples. Report skipped optional-backend\u002Fnetwork checks rather than claiming\n   support for them.\n\n## Common statistical recipes\n\n```python\nimport torch\n\nfrom earth2studio.statistics import lat_weight, mean, rmse\n\nweights = lat_weight(torch.as_tensor(latitudes))\narea_mean = mean([\"lat\"], weights=weights)\nscore = rmse([\"lat\", \"lon\"], weights=grid_weights,\n             ensemble_dimension=\"ensemble\")\n```\n\n`mean`, `variance`, and `std` can maintain state with `batch_update=True`.\nMetrics such as RMSE\u002FMAE can reduce an ensemble mean when\n`ensemble_dimension` is supplied. CRPS, rank histograms, and energy score expect\nan ensemble axis in forecast `x` and no such axis in truth `y`. Use the exact\nweight dimensionality and output-coordinate rules in the statistics reference.\n\n## Safe validation commands\n\nFrom any directory, run the bundled checker by its installed skill path:\n\n```bash\nuv run python \u003Cskill-root>\u002Fscripts\u002Fcontract_smoke.py --help\nuv run python \u003Cskill-root>\u002Fscripts\u002Fcontract_smoke.py --case all\n```\n\nWhen developing a custom component in your own project, run that project's\nfocused tests in its managed environment after the bundled contract smoke. A\nmissing optional Earth2Studio group is a limitation to report, not a reason to\ninstall an unbounded dependency set. See the testing reference for the local\nfixture expectations; the original repository's native test paths are review\nevidence, not runtime dependencies.\n\n## Scope boundaries and handoff\n\nThis sub-skill covers component implementation, statistics\u002Fmetric axes and\nweights, lexicon\u002Ftime-tolerance conventions, focused tests, and safe local\nvalidation. It intentionally excludes release\u002FCI operations, package-wide\nmaintainer workflows, serving-client construction, exhaustive catalogues, and\nremote data\u002Fmodel downloads. Preserve those limits in the downstream handoff.\nRecord native candidates and the two difficult synthetic cases: a diagnostic\nwhose returned coordinate mapping omits its output `variable`, and an ensemble\nmetric whose truth axes are reordered but retain a broadcastable shape.\n",{"name":5,"license":23,"metadata":24,"description":7,"disable-model-invocation":26},"Apache 2.0",{"disco-role":25},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension",27,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab","2026-09-03 10:25:21",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":26},[36,39,42],{"tool":37,"command":38},"npx","npx skills add VectorSpaceLab\u002FAREX-Skill --skill statistics-and-extension",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill statistics-and-extension",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill statistics-and-extension","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002FSKILL.md",[47,53,57,61,65],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Fextension-reference.md","REFERENCE","参考文档",9104,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002Freferences\u002Fextension-reference.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Fstatistics-reference.md",7397,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002Freferences\u002Fstatistics-reference.md",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Ftesting-reference.md",6756,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002Freferences\u002Ftesting-reference.md",{"path":62,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Ftroubleshooting.md",6846,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002Freferences\u002Ftroubleshooting.md",{"path":66,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":69,"sourceHtmlUrl":70},"scripts\u002Fcontract_smoke.py","SCRIPT","脚本",1695,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fearth2studio\u002Fsub-skills\u002Fstatistics-and-extension\u002Fscripts\u002Fcontract_smoke.py",[72,79,86,93,100,107],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":18},"01M799NYSH1R3RVAHSM2D88Y84","passnet-triton-opt","Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block\u002Fgrid\u002Fwarp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-triton-opt\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":18},"01M799NS5APEC0N1FZ7XZVW6BN","passnet-solve","END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-solve\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":18},"01M799NNY9BKSM21V5R7QVTBJX","passnet-skill","PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under .\u002Fpass_dir\u002F, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-skill\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":18},"01M799NHH6NP7Z1RK77KGXAN6S","passnet-pattern-fusion","Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()\u002Freplacement_args()\u002Freplacement_func() files, when a pass \"failed to match\", or when deciding how to fuse multiple kernels\u002Fops.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-pattern-fusion\u002FSKILL.md",[],[],{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":18},"01M799ND2J4WNC2PGAZ0A06NDF","passnet-orchestrate","Round\u002Fbatch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep\u002Frevert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-orchestrate\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":113,"publishedAt":18},"01M799N8SBTMY02ZDG2AZF5ZWE","passnet-feedback","Fast, deterministic feedback for PassNet work: pre-flight pattern\u002Fmatch verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-feedback\u002FSKILL.md",[],[]]