TorchedHat
Official@torchedhat
Offers specialized diagnostic and development capabilities for PyTorch compilation pipelines, distributed training stability, and custom configuration definition files.
Agent Skills by TorchedHat
Showing 16 vetted skills indexed across 1 GitHub repositories.
TorchedHattest-refactor
Refactor PyTorch test files to classify test classes and remove hardware-specific assumptions.
TorchedHatrun-inside-test
Executes a bundled test script and reports its exact output.
TorchedHatdistributed-hang-diagnosis
Diagnose PyTorch distributed training hangs and identify root causes.
TorchedHatagent-writer
Create and update Claude Code agent definition files with YAML frontmatter and validation.
TorchedHatplugin-writer
Create and configure Claude Code plugins with valid metadata and directory structures.
TorchedHatskill-writer
Create and validate Claude Code Agent Skill definitions with SKILL.md structure.
TorchedHatskill-developer
Create and manage Claude Code skills with YAML frontmatter and skill-rules.json triggers.
TorchedHatpytorch-dynamo
Debug and extend PyTorch Dynamo compilation across symbolic execution and FX graph generation.
TorchedHatcompile-trace-dynamo
Diagnose PyTorch Dynamo compilation failures via FX graph inspection.
TorchedHatpytorch-inductor
Diagnose and optimize PyTorch TorchInductor compilation behavior and generated kernels.
TorchedHatcompile-trace-inductor
Diagnose PyTorch Inductor compilation and performance issues by tracing IR transformations.
TorchedHatpytorch-aot
Guide PyTorch functorch and AOT Autograd implementation and extension workflows.
TorchedHatcompile-bisect
Bisect PyTorch compiler stages to isolate failing backend, subsystem, and operation.
TorchedHatcompile-overview
Explain torch.compile pipeline stages and intermediate representations for debugging.
TorchedHatcompile-trace-aot
Trace PyTorch AOT Autograd FX graphs to debug compiled training workflows.
TorchedHatvllm-compile
Diagnoses and explains JetBrains Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey Odyssey issues.
Frequently Asked Questions About TorchedHat
FAQPage SchemaWhat specific tasks can be performed using TorchedHat's diagnostic capabilities?▼
These capabilities enable deep inspection of PyTorch compilation stages, including FX graph analysis, Inductor kernel optimization, and AOT Autograd tracing. Additionally, users can perform distributed training hang diagnosis and manage structured definition files for custom environment extensions.
Which technical personas benefit most from these diagnostic skills?▼
Machine learning engineers, compiler researchers, and infrastructure developers working on high-performance model training benefit from these skills. They are specifically designed for those debugging complex PyTorch compilation failures or optimizing distributed training performance across heterogeneous hardware clusters.
What are the prerequisites for utilizing these compilation and diagnostic skills?▼
Users require an existing PyTorch environment with access to the internal compilation stack, including Dynamo and Inductor. Familiarity with FX graph representations and intermediate IR transformations is necessary to interpret the diagnostic output provided by these modules.