[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79148VM40JXQTCTBESH6AXJ":3,"skill-same:01M79148VM40JXQTCTBESH6AXJ":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},"01M79148VM40JXQTCTBESH6AXJ","speechlm2",null,"Use for NeMo SpeechLM2\u002FSALM\u002FSALMAutomodel, duplex STT\u002FS2S\u002FEAR-TTS, Nemotron VoiceChat, HuggingFace export\u002Fconversion, vLLM plugin, Automodel backends, and voice-agent troubleshooting.","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\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: speechlm2\ndescription: \"Use for NeMo SpeechLM2\u002FSALM\u002FSALMAutomodel, duplex STT\u002FS2S\u002FEAR-TTS,\n  Nemotron VoiceChat, HuggingFace export\u002Fconversion, vLLM plugin, Automodel\n  backends, and voice-agent troubleshooting.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# SpeechLM2\n\nUse this sub-skill for NeMo SpeechLM2 speech-LLM work: SALM\u002FSALMAutomodel, duplex STT and speech-to-speech models, Duplex EAR-TTS inside SpeechLM2 pipelines, Nemotron VoiceChat, HuggingFace\u002Fsafetensors export, the NeMo SpeechLM vLLM plugin, Automodel optional compiled backends, and local voice-agent configuration.\n\nThis guidance is self-contained. It distills repository evidence from `docs\u002Fsource\u002Fspeechlm2\u002Fintro.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fmodels.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fdatasets.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fconfigs.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Ftraining_and_scaling.rst`, `docs\u002Fsource\u002Fcheckpoints\u002Fintro.rst`, `examples\u002Fspeechlm2\u002F*.py`, `examples\u002Fspeechlm2\u002Fconf\u002F*.yaml`, `examples\u002Fvoice_agent\u002F**`, `nemo\u002Fcollections\u002Fspeechlm2\u002F**`, `nemo\u002Fagents\u002Fvoice_agent\u002F**`, and `tests\u002Fcollections\u002Fspeechlm2\u002F**`.\n\n## Route Here When\n\n- The task mentions `nemo.collections.speechlm2`, `SALM`, `SALMAutomodel`, `SALMWithAsrDecoder`, `DuplexS2SModel`, `DuplexS2SSpeechDecoderModel`, `DuplexSTTModel`, `DuplexEARTTS`, or `NemotronVoiceChat`.\n- The user needs speech+LLM generation, SpeechLM2 training configs, Lhotse conversation datasets, duplex speech-to-speech\u002FSTT evaluation, or SpeechLM2 inference with audio placeholder tokens.\n- The task involves SpeechLM2 HuggingFace export, `model.safetensors`, distributed checkpoint consolidation, `NemotronVoiceChat.from_pretrained`, or vLLM serving with `nemo_speechlm`.\n- The task involves NeMo voice-agent server\u002Fclient configuration, tool calling, vLLM\u002FHF LLM backend selection, streaming ASR\u002FTTS service composition, or backchannel\u002Fturn-taking settings.\n\n## Route Elsewhere\n\n- Classic ASR transcription, ASR fine-tuning, CTC\u002FRNNT decoding, Parakeet\u002FCanary model recipes, or ASR manifests belong in `..\u002Fasr\u002FSKILL.md` unless SpeechLM2 is explicitly involved.\n- TTS-only FastPitch, HiFiGAN, Magpie, Kokoro, or speech synthesis outside a SpeechLM2\u002FNemotron VoiceChat\u002Fvoice-agent pipeline belongs in `..\u002Ftts\u002FSKILL.md`.\n- Repo development, test authoring, packaging, lint, CI, and source-tree maintenance belong in the repo-development sub-skill when available.\n\n## First Steps\n\n1. Identify the workflow: SALM generation\u002Feval, SALM\u002FSALMAutomodel training, duplex STT\u002FS2S, Duplex EAR-TTS inference, Nemotron VoiceChat, HF export, vLLM serving, or voice-agent deployment.\n2. Read `references\u002Fworkflows.md` for the concrete API\u002Fcommand-shape workflow and safe adaptation notes.\n3. Read `references\u002Fconfiguration.md` before editing YAML, Hydra overrides, dataset paths, conversation roles, parallelism, or voice-agent config.\n4. Read `references\u002Fcheckpoints-and-export.md` before loading, resuming, converting, exporting, or serving checkpoints.\n5. Read `references\u002Ftroubleshooting.md` when imports, CUDA, compiled backends, data schemas, Hydra overrides, vLLM, or voice-agent services fail.\n6. Run `scripts\u002Fcheck_speechlm2_config.py --help` and then use it on user-provided SpeechLM2 YAML\u002FJSON configs before long-running jobs.\n\n## Core Facts\n\n- SpeechLM2 is active development code for augmenting pretrained LLMs with speech understanding and speech generation; expect APIs\u002Fconfigs to change more often than classic ASR\u002FTTS.\n- `SALM` uses HuggingFace Transformers for the LLM backbone; `SALMAutomodel` uses NeMo Automodel and is the right path for FSDP2, TP, PP, CP, EP\u002FMoE, HSDP, native Automodel LoRA, and distributed inference.\n- Distributed SALM generation\u002Fevaluation requires `SALMAutomodel`; plain `SALM` should stay single-process\u002Fsingle-device for inference.\n- `NemotronVoiceChat` is inference-only. Train or fine-tune its STT and EAR-TTS components separately, then export a joint VoiceChat checkpoint.\n- SpeechLM2 model checkpoints are HuggingFace-style directories with `config.json` and `model.safetensors`; `.nemo` is used in this collection mainly as an input format for pretrained ASR\u002Fperception components.\n- Training and most inference require a compatible GPU\u002FCUDA\u002FPyTorch stack. Optional Automodel compiled backends such as TransformerEngine, DeepEP, grouped GEMM, flash-attn, vLLM, Pipecat, browser\u002Fclient tooling, and Node\u002Fnpm must be installed only when that workflow actually needs them.\n\n## Bundled Assets\n\n- `references\u002Fworkflows.md`: SALM\u002FSALMAutomodel generation and training, duplex STT\u002FS2S\u002FEAR-TTS, Nemotron VoiceChat, vLLM, and voice-agent workflows.\n- `references\u002Fconfiguration.md`: SpeechLM2 YAML structure, data schemas, Hydra overrides, parallelism, and voice-agent configuration patterns.\n- `references\u002Fcheckpoints-and-export.md`: checkpoint formats, `from_pretrained`, `save_pretrained`, HF export, vLLM readiness, and Nemotron VoiceChat conversion.\n- `references\u002Ftroubleshooting.md`: actionable failure diagnosis for install\u002Fimport, CUDA\u002Fbackends, data\u002Fconfig, API\u002FCLI, export, vLLM, and voice-agent issues.\n- `scripts\u002Fcheck_speechlm2_config.py`: deterministic local checker for SpeechLM2 YAML\u002FJSON-like configs; it performs text-level validation without importing NeMo or downloading models.\n","# SpeechLM2\n\nUse this sub-skill for NeMo SpeechLM2 speech-LLM work: SALM\u002FSALMAutomodel, duplex STT and speech-to-speech models, Duplex EAR-TTS inside SpeechLM2 pipelines, Nemotron VoiceChat, HuggingFace\u002Fsafetensors export, the NeMo SpeechLM vLLM plugin, Automodel optional compiled backends, and local voice-agent configuration.\n\nThis guidance is self-contained. It distills repository evidence from `docs\u002Fsource\u002Fspeechlm2\u002Fintro.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fmodels.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fdatasets.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Fconfigs.rst`, `docs\u002Fsource\u002Fspeechlm2\u002Ftraining_and_scaling.rst`, `docs\u002Fsource\u002Fcheckpoints\u002Fintro.rst`, `examples\u002Fspeechlm2\u002F*.py`, `examples\u002Fspeechlm2\u002Fconf\u002F*.yaml`, `examples\u002Fvoice_agent\u002F**`, `nemo\u002Fcollections\u002Fspeechlm2\u002F**`, `nemo\u002Fagents\u002Fvoice_agent\u002F**`, and `tests\u002Fcollections\u002Fspeechlm2\u002F**`.\n\n## Route Here When\n\n- The task mentions `nemo.collections.speechlm2`, `SALM`, `SALMAutomodel`, `SALMWithAsrDecoder`, `DuplexS2SModel`, `DuplexS2SSpeechDecoderModel`, `DuplexSTTModel`, `DuplexEARTTS`, or `NemotronVoiceChat`.\n- The user needs speech+LLM generation, SpeechLM2 training configs, Lhotse conversation datasets, duplex speech-to-speech\u002FSTT evaluation, or SpeechLM2 inference with audio placeholder tokens.\n- The task involves SpeechLM2 HuggingFace export, `model.safetensors`, distributed checkpoint consolidation, `NemotronVoiceChat.from_pretrained`, or vLLM serving with `nemo_speechlm`.\n- The task involves NeMo voice-agent server\u002Fclient configuration, tool calling, vLLM\u002FHF LLM backend selection, streaming ASR\u002FTTS service composition, or backchannel\u002Fturn-taking settings.\n\n## Route Elsewhere\n\n- Classic ASR transcription, ASR fine-tuning, CTC\u002FRNNT decoding, Parakeet\u002FCanary model recipes, or ASR manifests belong in `..\u002Fasr\u002FSKILL.md` unless SpeechLM2 is explicitly involved.\n- TTS-only FastPitch, HiFiGAN, Magpie, Kokoro, or speech synthesis outside a SpeechLM2\u002FNemotron VoiceChat\u002Fvoice-agent pipeline belongs in `..\u002Ftts\u002FSKILL.md`.\n- Repo development, test authoring, packaging, lint, CI, and source-tree maintenance belong in the repo-development sub-skill when available.\n\n## First Steps\n\n1. Identify the workflow: SALM generation\u002Feval, SALM\u002FSALMAutomodel training, duplex STT\u002FS2S, Duplex EAR-TTS inference, Nemotron VoiceChat, HF export, vLLM serving, or voice-agent deployment.\n2. Read `references\u002Fworkflows.md` for the concrete API\u002Fcommand-shape workflow and safe adaptation notes.\n3. Read `references\u002Fconfiguration.md` before editing YAML, Hydra overrides, dataset paths, conversation roles, parallelism, or voice-agent config.\n4. Read `references\u002Fcheckpoints-and-export.md` before loading, resuming, converting, exporting, or serving checkpoints.\n5. Read `references\u002Ftroubleshooting.md` when imports, CUDA, compiled backends, data schemas, Hydra overrides, vLLM, or voice-agent services fail.\n6. Run `scripts\u002Fcheck_speechlm2_config.py --help` and then use it on user-provided SpeechLM2 YAML\u002FJSON configs before long-running jobs.\n\n## Core Facts\n\n- SpeechLM2 is active development code for augmenting pretrained LLMs with speech understanding and speech generation; expect APIs\u002Fconfigs to change more often than classic ASR\u002FTTS.\n- `SALM` uses HuggingFace Transformers for the LLM backbone; `SALMAutomodel` uses NeMo Automodel and is the right path for FSDP2, TP, PP, CP, EP\u002FMoE, HSDP, native Automodel LoRA, and distributed inference.\n- Distributed SALM generation\u002Fevaluation requires `SALMAutomodel`; plain `SALM` should stay single-process\u002Fsingle-device for inference.\n- `NemotronVoiceChat` is inference-only. Train or fine-tune its STT and EAR-TTS components separately, then export a joint VoiceChat checkpoint.\n- SpeechLM2 model checkpoints are HuggingFace-style directories with `config.json` and `model.safetensors`; `.nemo` is used in this collection mainly as an input format for pretrained ASR\u002Fperception components.\n- Training and most inference require a compatible GPU\u002FCUDA\u002FPyTorch stack. Optional Automodel compiled backends such as TransformerEngine, DeepEP, grouped GEMM, flash-attn, vLLM, Pipecat, browser\u002Fclient tooling, and Node\u002Fnpm must be installed only when that workflow actually needs them.\n\n## Bundled Assets\n\n- `references\u002Fworkflows.md`: SALM\u002FSALMAutomodel generation and training, duplex STT\u002FS2S\u002FEAR-TTS, Nemotron VoiceChat, vLLM, and voice-agent workflows.\n- `references\u002Fconfiguration.md`: SpeechLM2 YAML structure, data schemas, Hydra overrides, parallelism, and voice-agent configuration patterns.\n- `references\u002Fcheckpoints-and-export.md`: checkpoint formats, `from_pretrained`, `save_pretrained`, HF export, vLLM readiness, and Nemotron VoiceChat conversion.\n- `references\u002Ftroubleshooting.md`: actionable failure diagnosis for install\u002Fimport, CUDA\u002Fbackends, data\u002Fconfig, API\u002FCLI, export, vLLM, and voice-agent issues.\n- `scripts\u002Fcheck_speechlm2_config.py`: deterministic local checker for SpeechLM2 YAML\u002FJSON-like configs; it performs text-level validation without importing NeMo or downloading models.\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\u002Fnemo\u002Fsub-skills\u002Fspeechlm2",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 speechlm2",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill speechlm2",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill speechlm2","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002FSKILL.md",[47,53,57,61,65],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Fcheckpoints-and-export.md","REFERENCE","参考文档",10316,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002Freferences\u002Fcheckpoints-and-export.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Fconfiguration.md",11904,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002Freferences\u002Fconfiguration.md",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Ftroubleshooting.md",11089,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002Freferences\u002Ftroubleshooting.md",{"path":62,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Fworkflows.md",12671,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002Freferences\u002Fworkflows.md",{"path":66,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":69,"sourceHtmlUrl":70},"scripts\u002Fcheck_speechlm2_config.py","SCRIPT","脚本",12981,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fnemo\u002Fsub-skills\u002Fspeechlm2\u002Fscripts\u002Fcheck_speechlm2_config.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",[],[]]