chisuhua
Community@chisuhua
chisuhua publishes 22 skills centered on C++ modernization, CUDA kernel optimization, CMake builds, debugging, and web research via Tavily and inference.sh.
Agent Skills by chisuhua
Showing 20 vetted skills indexed across 1 GitHub repositories.
chisuhuaplanning-with-files
Creates persistent markdown planning files to track phases, findings, and progress across sessions.
chisuhuacpp-debug
Diagnose C++ runtime failures including segfaults, deadlocks, and memory leaks.
chisuhuacpp-modernize
Modernizes legacy C++98/03 code to C++17/20 with smart pointers and nullptr.
chisuhuacpp-architecture
Analyzes C++ project dependencies and module boundaries using clang AST and Mermaid diagrams.
chisuhuaobsidian-markdown
Create and edit Obsidian Flavored Markdown with wikilinks, embeds, callouts, and properties.
chisuhuacmake-manage
Manages CMake build configurations, dependencies, presets, and cross-platform setups.
chisuhuacuda
Debug, profile, and optimize CUDA kernels using nsys, ncu, compute-sanitizer, and PTX documentation.
chisuhuastate-modification-audit
Audits all read and write paths of a state variable to trace unexpected modifications.
chisuhuaregression-bisect
Diagnose test regressions after refactoring using git bisect and semantic diff analysis.
chisuhuaagent-browser
Automate browser navigation, element interaction, and content extraction via Playwright sessions.
chisuhuaagent-tools
Run 150+ cloud AI apps for image, video, LLM, search, and Twitter tasks via the inference.sh CLI.
chisuhuacmake
Write and configure modern target-based CMakeLists.txt for C/C++ projects.
chisuhuagog
Manages Gmail, Calendar, Drive, Contacts, Sheets, and Docs via the gog command-line interface.
chisuhuacpp-pro
Writes, optimizes, and debugs C++ code using C++20/23 features, templates, and CMake tooling.
chisuhuacuda-kernels
Write and benchmark optimized CUDA kernels for HuggingFace diffusers and transformers models.
chisuhuatavily-research
Generates cited research reports from web sources via the Tavily CLI.
chisuhuatavily-search
Search the web via the Tavily CLI and return LLM-optimized results with snippets and relevance scores.
chisuhuafind-skills
Searches and installs agent skills from the open skills ecosystem using the Skills CLI.
chisuhuaweb-search
Search the web and extract page content via Tavily and Exa through the inference.sh CLI.
chisuhuacouncil-workflow
Orchestrates multi-agent PRD and ADR creation with resumable document-driven state tracking.
Frequently Asked Questions About chisuhua
FAQPage SchemaWhat tasks can I accomplish with chisuhua's skills?▼
You can modernize legacy C++ to C++17/20, debug segfaults and memory leaks, analyze C++ architecture dependencies, manage CMake builds and cross-platform presets, write and benchmark CUDA kernels for H100/A100 GPUs, audit state-variable modifications, bisect regressions, and run cited web research via Tavily and inference.sh.
Who are these skills designed for?▼
Systems and C++ engineers maintaining legacy codebases, GPU developers optimizing kernels for HuggingFace diffusers and transformers models like LLaMA and Stable Diffusion, and developers needing structured multi-step task planning, Obsidian markdown editing, or Google Workspace operations via the gog CLI.
What prerequisites and dependencies do these skills require?▼
cpp-modernize requires clang-tidy, cmake, and a C++17+ compiler. CUDA skills need NVIDIA profiling tools (nsys, ncu, cuda-gdb, compute-sanitizer). Tavily and inference.sh skills require the tvly and infsh CLIs. The gog skill installs via Homebrew (steipete/tap/gogcli). regression-bisect depends on a git-master skill.
Are these skills open source and what licenses apply?▼
cpp-modernize is licensed Apache-2.0 and cpp-pro is MIT licensed, permitting free use and modification. Other skills in the manifest do not declare explicit licenses, so users should check individual repository folders before redistribution or commercial use.
How do the planning and research skills work in practice?▼
planning-with-files creates task_plan.md, findings.md, and progress.md for complex multi-step tasks, with hooks that reload context after session clears. tavily-research produces cited multi-source reports in 30-120 seconds, while tavily-search returns quick LLM-optimized results with relevance scores and domain filtering.