[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6PSBYF3QVADR1TBBEC5XRPE":3,"skill-same:01M6PSBYF3QVADR1TBBEC5XRPE":157},{"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":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":59,"commitSha":60,"skillDirectory":61,"forksCount":14,"repositoryHtmlUrl":62,"repositoryHomepageUrl":6,"ownerHtmlUrl":63,"ownerBlogUrl":64,"ownerDescription":6,"updatedAt":65,"install":66,"resources":79},"01M6PSBYF3QVADR1TBBEC5XRPE","tao-finetune-cosmos-reason",null,"Cosmos-Reason2-8B video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video question-answering models, fine-tuning Cosmos-Reason2 with SFT, or working with Cosmos-RL. Trigger phrases include \"fine-tune Cosmos-Reason\", \"Cosmos-RL SFT\", \"video QA fine-tune\", \"Cosmos-Reason2-8B training\".","headmktlucca-ui\u002Fneuroads-lp","headmktlucca-ui","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F247379838?v=4","01M6PDVAV6M1W96D1SR1PDF8CS","NeuroAds","01M6PDVAZ4P281GNDKD45CYTF5",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"tao finetune cosmos reason","tao","reason","finetune","cosmos","container","cosmos-reason","FSDP","trigger","rl","docker","use","question","Bash","Video","Fine-Tuning","NVIDIA","qa","toolkit","training","2026-11-10 10:12:30",false,"---\nname: tao-finetune-cosmos-reason\ndescription: Cosmos-Reason2-8B video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video\n  question-answering models, fine-tuning Cosmos-Reason2 with SFT, or working with Cosmos-RL. Trigger phrases include\n  \"fine-tune Cosmos-Reason\", \"Cosmos-RL SFT\", \"video QA fine-tune\", \"Cosmos-Reason2-8B training\".\nlicense: Apache-2.0\ncompatibility: Requires docker + nvidia-container-toolkit.\nmetadata:\n  author: NVIDIA Corporation\n  version: \"0.1.0\"\nallowed-tools: Read Bash\ntags:\n- video\n- qa\n- cosmos\n- sft\n- reasoning\n- vlm\n---\n\n# Cosmos-RL\n\nSupervised fine-tuning (SFT) of **nvidia\u002FCosmos-Reason2-8B** on video reasoning tasks. Pretrained weights are sourced from HuggingFace, not NGC. This is a **gated model** — requires `HF_TOKEN`.\n\nUses FSDP-based parallelism with `dp_shard_size` for GPU count and `dp_replicate_size` for node count (not the standard `num_gpus`\u002F`num_nodes`).\n\n## When to Use\n\nUse this skill to train, evaluate, quantize, or run inference on Cosmos-Reason2-8B for video question-answering and video reasoning. The core workflow is: confirm `HF_TOKEN` gating, sample annotations for `video_fps`, load the spec template, apply the critical train overrides below, then launch through the platform skill (or AutoML when enabled).\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only; otherwise default to `auto`. When `automl_policy: auto`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Credentials\n\n- **HF_TOKEN** (required): HuggingFace access token. The user must accept the model agreement at \u003Chttps:\u002F\u002Fhuggingface.co\u002Fnvidia\u002FCosmos-Reason2-8B> and provide a token with read access. Passed to the container as a `docker_env_var`.\n\n## Datasets\n\nDataset type is **vlm** in **llava** format; accepted intents are training, evaluation, and testing. Inputs may be dataset roots (root mode maps `\u003Croot>\u002Fannotations.json` plus `\u003Croot>` as the media path) or direct spec-key paths (when annotations and media live in different locations). Before launching train\u002FAutoML\u002Fevaluate, sample the annotation JSON and require `video_fps` in each record — missing `video_fps` makes the Cosmos-RL SFT loader fail with `Error processing sample: 'video_fps'` after the job starts. Stop before runner generation if it is absent and ask the user to fix the annotation files; do not start AutoML to discover this inside torchrun.\n\nSee `references\u002Fdatasets.md` for the full training requirements, the launch intake reminder (spec-key options, root-mode mapping, container-image confirmation, and the `check_tao_launch_preflight.py` invocation), the Per-Action Dataset Requirements table, the `data_sources` mapping with direct-override examples, and the eval-dataset \u002F auto-split policy.\n\n## Spec Construction\n\ncosmos-rl is `mode: config`. **Always start from `references\u002Fspec_template_train.yaml`** (or `spec_template_evaluate.yaml` for evaluate) — load it via `yaml.safe_load(...)` and apply user overrides on top. The spec the model consumes is **nested dicts**, not flat dotted keys; the dotted override notation denotes paths into the nested spec, so walk the path and assign at the leaf. Data source overrides are **mandatory for every action** and must be built from the Per-Action Dataset Requirements table in `references\u002Fdatasets.md`.\n\nSee `references\u002Fspec-construction.md` for the load-template-then-override pattern and the full typical override blocks for train (including `policy.model_max_length=81920`, `dp_shard_size`\u002F`dp_replicate_size`, and LoRA `lora_alpha`\u002F`r`\u002F`lora_dropout`), evaluate, quantize, and inference, plus the note that `custom.val_dataset` leaf keys are valid even when absent from the default spec object.\n\n## Critical Overrides (Train)\n\nThese are the keys whose template defaults are wrong or where omission flips the run into a different mode:\n\n| Parameter | Template Default | Required Value | Why |\n|---|---|---|---|\n| `policy.model_name_or_path` | `nvidia\u002FCosmos-Reason2-8B` | `hf_model:\u002F\u002Fnvidia\u002FCosmos-Reason2-8B` (or local checkpoint) | The bare HF id makes cosmos-rl fetch from HF Hub at runtime; the `hf_model:\u002F\u002F` URI form pre-downloads the weights before the training command starts |\n| `policy.model_max_length` | 40960 | Keep at 40960 or higher | Smaller than ~40k causes `vision_embeds` shape mismatch on video inputs |\n| `train.train_batch_per_replica` | 32 | Any multiple of `train.train_policy.mini_batch` | Mismatch raises an immediate AssertionError |\n| `train.train_policy.type` | `\"sft\"` | Keep as `\"sft\"` for SFT workflows | If dropped during agent regeneration, cosmos-rl flips to RL mode → rollout replica allocated → multi-node attempted → hostname errors when `num_nodes=1` |\n\n## Parameters\n\n`train.train_batch_per_replica` must be divisible by `train.train_policy.mini_batch`; `policy.model_max_length` must be 40960 or higher for video SFT; `policy.parallelism.dp_shard_size` should equal GPUs per node and `dp_replicate_size` the node count; `custom.vision.fps` and `custom.vision.nframes` are mutually exclusive (set exactly one). Cosmos-RL models are 8B parameters and benefit from multi-GPU FSDP sharding — recommended: 8x A100 or H100 (80GB each).\n\nSee `references\u002Fparameters.md` for the complete parameter reference: training loop, model & policy, parallelism (including multi-node guidance and platform-skill pointers), optimization & data loading, vision encoders (fps vs nframes details and the decord\u002Ftorchvision failure mode), checkpointing, validation, logging, and hardware.\n\n## Evaluate\n\nThe evaluator reads a **flat TOML** config with top-level keys `dataset`, `model`, `task`, `evaluation`, `vision`, `generation`, `metrics`, `results`, `num_gpus`, `results_dir`. Task type is `\"\"` (General Evaluator, auto-detects binary yes\u002Fno classification and computes TP\u002FFP\u002FTN\u002FFN\u002Faccuracy\u002Fprecision\u002Frecall\u002FF1) or `\"its_directionality\"` (left\u002Fright\u002Fstraight; do NOT use for collision detection). The `actions.evaluate` block in `references\u002Fskill_info.yaml` declares inputs and outputs; for SDK invocation see `skills\u002Fplatform\u002Ftao-run-platform\u002FSKILL.md`.\n\nSee `references\u002Fevaluate.md` for the config-format detail, task-type notes, LoRA evaluation (checkpoint path via `spec_overrides` with `model.enable_lora`\u002F`model.base_model_path` and adapter merge behavior), selective download (`{annotation, format, keys}` partial media pull), and the results format and metrics.\n\n## Error Patterns\n\nCommon failures include CUDA OOM in train (reduce `mini_batch` or raise `dp_shard_size`), OOM during LoRA evaluation, NaN loss, the `vision_embeds` shape mismatch (raise `model_max_length` to 40960), `train_batch_per_replica` not divisible by `mini_batch`, `train_batch_per_replica` larger than samples per rank (the `'NoneType' object has no attribute 'state_dict'` 0-step crash), stale dataset cache after changing fps\u002Ftotal_pixels, and the gated-repo authentication loop.\n\nSee `references\u002Ftroubleshooting.md` for the full diagnosis and fix for each error pattern.\n\n## DEFT Support and Parent-Model Inference\n\nCosmos-RL implements the DEFT workflow contract for video QA tasks (see `config.json` and `workflow\u002Fdeft\u002Fdeft.md`). Gap analysis via `scripts\u002Fanalyze_gaps.py` reads cosmos-rl `results.json`, compares predictions by exact string match after `.lower().strip()`, and emits a parquet of failure cases — so eval prompts must force short constrained answers. Model-specific parent-model inference mappings (evaluate\u002Finference\u002Fquantize\u002Ftrain spec fields → inference functions, checkpoint metadata, and `parent_job_id` handling) live in the reference, not in `config.json`.\n\nSee `references\u002Fdeft-and-inference-mappings.md` for the gap-analysis detail and limitation, and the full parent-model inference mapping table.\n","# Cosmos-RL\n\nSupervised fine-tuning (SFT) of **nvidia\u002FCosmos-Reason2-8B** on video reasoning tasks. Pretrained weights are sourced from HuggingFace, not NGC. This is a **gated model** — requires `HF_TOKEN`.\n\nUses FSDP-based parallelism with `dp_shard_size` for GPU count and `dp_replicate_size` for node count (not the standard `num_gpus`\u002F`num_nodes`).\n\n## When to Use\n\nUse this skill to train, evaluate, quantize, or run inference on Cosmos-Reason2-8B for video question-answering and video reasoning. The core workflow is: confirm `HF_TOKEN` gating, sample annotations for `video_fps`, load the spec template, apply the critical train overrides below, then launch through the platform skill (or AutoML when enabled).\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only; otherwise default to `auto`. When `automl_policy: auto`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Credentials\n\n- **HF_TOKEN** (required): HuggingFace access token. The user must accept the model agreement at \u003Chttps:\u002F\u002Fhuggingface.co\u002Fnvidia\u002FCosmos-Reason2-8B> and provide a token with read access. Passed to the container as a `docker_env_var`.\n\n## Datasets\n\nDataset type is **vlm** in **llava** format; accepted intents are training, evaluation, and testing. Inputs may be dataset roots (root mode maps `\u003Croot>\u002Fannotations.json` plus `\u003Croot>` as the media path) or direct spec-key paths (when annotations and media live in different locations). Before launching train\u002FAutoML\u002Fevaluate, sample the annotation JSON and require `video_fps` in each record — missing `video_fps` makes the Cosmos-RL SFT loader fail with `Error processing sample: 'video_fps'` after the job starts. Stop before runner generation if it is absent and ask the user to fix the annotation files; do not start AutoML to discover this inside torchrun.\n\nSee `references\u002Fdatasets.md` for the full training requirements, the launch intake reminder (spec-key options, root-mode mapping, container-image confirmation, and the `check_tao_launch_preflight.py` invocation), the Per-Action Dataset Requirements table, the `data_sources` mapping with direct-override examples, and the eval-dataset \u002F auto-split policy.\n\n## Spec Construction\n\ncosmos-rl is `mode: config`. **Always start from `references\u002Fspec_template_train.yaml`** (or `spec_template_evaluate.yaml` for evaluate) — load it via `yaml.safe_load(...)` and apply user overrides on top. The spec the model consumes is **nested dicts**, not flat dotted keys; the dotted override notation denotes paths into the nested spec, so walk the path and assign at the leaf. Data source overrides are **mandatory for every action** and must be built from the Per-Action Dataset Requirements table in `references\u002Fdatasets.md`.\n\nSee `references\u002Fspec-construction.md` for the load-template-then-override pattern and the full typical override blocks for train (including `policy.model_max_length=81920`, `dp_shard_size`\u002F`dp_replicate_size`, and LoRA `lora_alpha`\u002F`r`\u002F`lora_dropout`), evaluate, quantize, and inference, plus the note that `custom.val_dataset` leaf keys are valid even when absent from the default spec object.\n\n## Critical Overrides (Train)\n\nThese are the keys whose template defaults are wrong or where omission flips the run into a different mode:\n\n| Parameter | Template Default | Required Value | Why |\n|---|---|---|---|\n| `policy.model_name_or_path` | `nvidia\u002FCosmos-Reason2-8B` | `hf_model:\u002F\u002Fnvidia\u002FCosmos-Reason2-8B` (or local checkpoint) | The bare HF id makes cosmos-rl fetch from HF Hub at runtime; the `hf_model:\u002F\u002F` URI form pre-downloads the weights before the training command starts |\n| `policy.model_max_length` | 40960 | Keep at 40960 or higher | Smaller than ~40k causes `vision_embeds` shape mismatch on video inputs |\n| `train.train_batch_per_replica` | 32 | Any multiple of `train.train_policy.mini_batch` | Mismatch raises an immediate AssertionError |\n| `train.train_policy.type` | `\"sft\"` | Keep as `\"sft\"` for SFT workflows | If dropped during agent regeneration, cosmos-rl flips to RL mode → rollout replica allocated → multi-node attempted → hostname errors when `num_nodes=1` |\n\n## Parameters\n\n`train.train_batch_per_replica` must be divisible by `train.train_policy.mini_batch`; `policy.model_max_length` must be 40960 or higher for video SFT; `policy.parallelism.dp_shard_size` should equal GPUs per node and `dp_replicate_size` the node count; `custom.vision.fps` and `custom.vision.nframes` are mutually exclusive (set exactly one). Cosmos-RL models are 8B parameters and benefit from multi-GPU FSDP sharding — recommended: 8x A100 or H100 (80GB each).\n\nSee `references\u002Fparameters.md` for the complete parameter reference: training loop, model & policy, parallelism (including multi-node guidance and platform-skill pointers), optimization & data loading, vision encoders (fps vs nframes details and the decord\u002Ftorchvision failure mode), checkpointing, validation, logging, and hardware.\n\n## Evaluate\n\nThe evaluator reads a **flat TOML** config with top-level keys `dataset`, `model`, `task`, `evaluation`, `vision`, `generation`, `metrics`, `results`, `num_gpus`, `results_dir`. Task type is `\"\"` (General Evaluator, auto-detects binary yes\u002Fno classification and computes TP\u002FFP\u002FTN\u002FFN\u002Faccuracy\u002Fprecision\u002Frecall\u002FF1) or `\"its_directionality\"` (left\u002Fright\u002Fstraight; do NOT use for collision detection). The `actions.evaluate` block in `references\u002Fskill_info.yaml` declares inputs and outputs; for SDK invocation see `skills\u002Fplatform\u002Ftao-run-platform\u002FSKILL.md`.\n\nSee `references\u002Fevaluate.md` for the config-format detail, task-type notes, LoRA evaluation (checkpoint path via `spec_overrides` with `model.enable_lora`\u002F`model.base_model_path` and adapter merge behavior), selective download (`{annotation, format, keys}` partial media pull), and the results format and metrics.\n\n## Error Patterns\n\nCommon failures include CUDA OOM in train (reduce `mini_batch` or raise `dp_shard_size`), OOM during LoRA evaluation, NaN loss, the `vision_embeds` shape mismatch (raise `model_max_length` to 40960), `train_batch_per_replica` not divisible by `mini_batch`, `train_batch_per_replica` larger than samples per rank (the `'NoneType' object has no attribute 'state_dict'` 0-step crash), stale dataset cache after changing fps\u002Ftotal_pixels, and the gated-repo authentication loop.\n\nSee `references\u002Ftroubleshooting.md` for the full diagnosis and fix for each error pattern.\n\n## DEFT Support and Parent-Model Inference\n\nCosmos-RL implements the DEFT workflow contract for video QA tasks (see `config.json` and `workflow\u002Fdeft\u002Fdeft.md`). Gap analysis via `scripts\u002Fanalyze_gaps.py` reads cosmos-rl `results.json`, compares predictions by exact string match after `.lower().strip()`, and emits a parquet of failure cases — so eval prompts must force short constrained answers. Model-specific parent-model inference mappings (evaluate\u002Finference\u002Fquantize\u002Ftrain spec fields → inference functions, checkpoint metadata, and `parent_job_id` handling) live in the reference, not in `config.json`.\n\nSee `references\u002Fdeft-and-inference-mappings.md` for the gap-analysis detail and limitation, and the full parent-model inference mapping table.\n",{"name":5,"tags":49,"license":15,"metadata":54,"description":7,"allowed-tools":57,"compatibility":58},[50,41,28,51,52,53],"video","sft","reasoning","vlm",{"author":55,"version":56},"NVIDIA Corporation","0.1.0","Read Bash","Requires docker + nvidia-container-toolkit.","main","7a940b8e5af16c95e5c190a524421d32fc7445a9",".claude\u002Fskills\u002Ftao-finetune-cosmos-reason","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui","https:\u002F\u002Fneuroads.com.br","2026-08-12 00:42:44",{"source":8,"skillName":5,"commands":67,"prompt":77,"zipAvailable":78},[68,71,74],{"tool":69,"command":70},"npx","npx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-finetune-cosmos-reason",{"tool":72,"command":73},"bunx","bunx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-finetune-cosmos-reason",{"tool":75,"command":76},"pnpm","pnpm dlx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-finetune-cosmos-reason","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002FSKILL.md",true,[80,86,90,94,98,102,106,112,116,120,124,128,132,136,140,144,147,151],{"path":81,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":84,"sourceHtmlUrl":85},"BENCHMARK.md","REFERENCE","参考文档",4700,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002FBENCHMARK.md",{"path":87,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":88,"sourceHtmlUrl":89},"evals\u002Fevals.json",821,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fevals\u002Fevals.json",{"path":91,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":92,"sourceHtmlUrl":93},"references\u002Fdatasets.md",4926,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fdatasets.md",{"path":95,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":96,"sourceHtmlUrl":97},"references\u002Fdeft-and-inference-mappings.md",2523,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fdeft-and-inference-mappings.md",{"path":99,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":100,"sourceHtmlUrl":101},"references\u002Fevaluate.md",2022,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fevaluate.md",{"path":103,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":104,"sourceHtmlUrl":105},"references\u002Fparameters.md",4273,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fparameters.md",{"path":107,"resourceType":108,"resourceTypeDesc":109,"sizeBytes":110,"sourceHtmlUrl":111},"references\u002Fscripts\u002Fanalyze_gaps.py","SCRIPT","脚本",4015,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fscripts\u002Fanalyze_gaps.py",{"path":113,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":114,"sourceHtmlUrl":115},"references\u002Fskill_info.yaml",2498,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fskill_info.yaml",{"path":117,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":118,"sourceHtmlUrl":119},"references\u002Fspec_template_evaluate.yaml",1183,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fspec_template_evaluate.yaml",{"path":121,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":122,"sourceHtmlUrl":123},"references\u002Fspec_template_train.yaml",1979,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fspec_template_train.yaml",{"path":125,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":126,"sourceHtmlUrl":127},"references\u002Fspec-construction.md",4852,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Fspec-construction.md",{"path":129,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":130,"sourceHtmlUrl":131},"references\u002Ftroubleshooting.md",2818,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Freferences\u002Ftroubleshooting.md",{"path":133,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":134,"sourceHtmlUrl":135},"schemas\u002Fevaluate.schema.json",18960,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fschemas\u002Fevaluate.schema.json",{"path":137,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":138,"sourceHtmlUrl":139},"schemas\u002Fmanifest.json",2025,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fschemas\u002Fmanifest.json",{"path":141,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":142,"sourceHtmlUrl":143},"schemas\u002Ftrain.schema.json",35966,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fschemas\u002Ftrain.schema.json",{"path":145,"resourceType":108,"resourceTypeDesc":109,"sizeBytes":110,"sourceHtmlUrl":146},"scripts\u002Fanalyze_gaps.py","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fscripts\u002Fanalyze_gaps.py",{"path":148,"resourceType":82,"resourceTypeDesc":83,"sizeBytes":149,"sourceHtmlUrl":150},"skill-card.md",3978,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fskill-card.md",{"path":152,"resourceType":153,"resourceTypeDesc":154,"sizeBytes":155,"sourceHtmlUrl":156},"skill.oms.sig","OTHER","其他",7989,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftao-finetune-cosmos-reason\u002Fskill.oms.sig",[158,182,207,233,251,286],{"skillNo":159,"name":160,"nameZh":6,"description":161,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":163,"categories":164,"keywords":171,"publishedAt":44},"01M6PTDCZTXTGSTX4RFK901YSZ","xunit-skill","Generates xUnit.net tests in C#. Covers Fact\u002FTheory, constructor injection, IClassFixture, and FluentAssertions. Use when user mentions \"xUnit\", \"[Fact]\", \"[Theory]\", \"Assert.Equal\", \"C# xUnit\". Triggers on: \"xUnit\", \"[Fact]\", \"[Theory]\", \"Assert.Equal C#\", \"xUnit.net\".","MIT","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fxunit-skill\u002FSKILL.md",[165,168],{"categoryNo":166,"categoryName":167,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":169,"categoryName":170,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[172,173,174,175,176,35,177,178,179,180,181],"xunit","xunit skill","testing","injection","user","tests","triggers","unit","unit-testing","C#",{"skillNo":183,"name":184,"nameZh":6,"description":185,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":186,"categories":187,"keywords":190,"publishedAt":44},"01M6PTCB59ZTG5FTMWGPYEMMBP","xcuitest-skill","Generates XCUITest UI tests for iOS\u002FiPadOS apps in Swift. Apple's native testing framework for reliable, fast UI automation. Supports local simulators and TestMu AI cloud real devices. Use when user mentions \"XCUITest\", \"XCTest\", \"iOS UI test\", \"Swift test\", \"XCUIApplication\". Triggers on: \"XCUITest\", \"XCTest UI\", \"iOS UI test\", \"Swift UI test\", \"XCUIApplication\", \"TestMu\".","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fxcuitest-skill\u002FSKILL.md",[188,189],{"categoryNo":166,"categoryName":167,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":169,"categoryName":170,"taxonomyType":21,"taxonomyTypeDesc":22},[191,192,174,193,194,195,196,197,35,198,199,200,201,202,177,178,203,204,205,206],"xcuitest","xcuitest skill","automation","test","testing framework","ios","devices","mobile","mobile-testing","ui test","local","apple","apps","native","for","objective",{"skillNo":208,"name":209,"nameZh":6,"description":210,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":211,"categories":212,"keywords":217,"publishedAt":44},"01M6PTATMNTY9NF4J7KSR76YK6","world-builder","Create and manage generative game zones for Dorothy's Pokemon-style world. Use this skill when creating, updating, or designing game worlds via MCP tools.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fworld-builder\u002FSKILL.md",[213,216],{"categoryNo":214,"categoryName":215,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":169,"categoryName":170,"taxonomyType":21,"taxonomyTypeDesc":22},[218,219,220,221,222,223,224,225,35,226,227,228,229,230,231,205,232],"builder","world","world builder","MCP","create","mcp tools","Tools","game","pokemon","manage","style","via","theme","zone","generative",{"skillNo":234,"name":235,"nameZh":6,"description":236,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":237,"categories":238,"keywords":241,"publishedAt":44},"01M6PT9Z92JTSQEHE1QR0NNW5F","webdriverio-skill","Generates WebdriverIO (WDIO) automation tests in JavaScript or TypeScript. Supports local and TestMu AI cloud. Use when user mentions \"WebdriverIO\", \"WDIO\", \"wdio.conf\", \"browser.url\", \"$\", \"$$\". Triggers on: \"WebdriverIO\", \"WDIO\", \"wdio\", \"browser.$\".","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fwebdriverio-skill\u002FSKILL.md",[239,240],{"categoryNo":166,"categoryName":167,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":169,"categoryName":170,"taxonomyType":21,"taxonomyTypeDesc":22},[242,174,243,193,244,245,246,247,176,35,248,201,177,178,249,250],"webdriverio skill","typescript","javascript","browser","cloud","e2e","e2e-testing","URL","conf",{"skillNo":252,"name":253,"nameZh":6,"description":254,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":255,"categories":256,"keywords":266,"publishedAt":44},"01M6PT9NVK36ZWCQ69YXK3FR4K","web-design-guidelines","Web platform design and accessibility guidelines. Use when building web interfaces, auditing accessibility, implementing responsive layouts, or reviewing web UI code. Triggers on tasks involving HTML, CSS, web components, WCAG compliance, responsive design, or web performance.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fweb\u002FSKILL.md",[257,260,261],{"categoryNo":258,"categoryName":259,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":214,"categoryName":215,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":262,"categoryName":263,"taxonomyType":264,"taxonomyTypeDesc":265},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[267,268,269,270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,178,285],"design","web","guidelines","web design guidelines","web design","css","performance","components","web components","tasks","code","web performance","ui","platform design","html","responsive design","responsive","platform","implementing",{"skillNo":287,"name":288,"nameZh":6,"description":289,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":162,"licenseFileUrl":6,"sourceHtmlUrl":290,"categories":291,"keywords":294,"publishedAt":44},"01M6PT9A3JKRE3X34276F2FW5R","watchos-design-guidelines","Apple Human Interface Guidelines for Apple Watch. Use when building watchOS apps, complications, or workout features. Triggers on tasks involving Watch UI, Digital Crown, glanceable interfaces, or wrist-based interactions.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fwatchos\u002FSKILL.md",[292,293],{"categoryNo":214,"categoryName":215,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":262,"categoryName":263,"taxonomyType":264,"taxonomyTypeDesc":265},[267,269,295,276,279,35,296,297,202,284,298,178,203,299,205,300,301],"watchos design guidelines","workout","human","watch","interface","platform-design","features"]