调试
定位并修复问题，分析错误
cuda
technillogue
"CUDA kernel development, debugging, and performance optimization for Claude Code. Use when writing, debugging, or optimizing CUDA code, GPU kernels, or parallel algorithms. Covers non-interactive profiling with nsys/ncu, debugging with cuda-gdb/compute-sanitizer, binary inspection with cuobjdump, and performance analysis workflows. Triggers on CUDA, GPU programming, kernel optimization, nsys, ncu, cuda-gdb, compute-sanitizer, PTX, GPU profiling, parallel performance."
ubs
Dicklesworthstone
"Ultimate Bug Scanner - Pre-commit static analysis for AI coding workflows. 18 detection categories, 8 languages, 4-layer analysis engine. The AI agent's quality gate."
inngest-middleware
inngest
Create and use Inngest middleware for cross-cutting concerns. Covers the middleware lifecycle, creating custom middleware, dependency injection with dependencyInjectionMiddleware, encryption via @inngest/middleware-encryption, Sentry error tracking via @inngest/middleware-sentry, and custom middleware patterns.
cleanup-notebooks
jsboige
Clean up and reorganize markdown in enriched Jupyter notebooks. Arguments: [target] [--dry-run] [--aggressive] [--hierarchy-only]
create-release
EdanStarfire
Create a versioned release of claudecode_webui — bump version fields, update CHANGELOG.md, tag the commit, push, and create the GitHub Release.
execute-notebook
jsboige
Execute Jupyter notebooks using local scripts or MCP Jupyter. Arguments: <notebook_path> [--mode] [--cells] [--timeout] [--batch] [--save]
model-router
oyi77
Intelligent model routing via subagents - automatically spawn the right model for each task complexity
axiom-sre
axiomhq
Expert SRE investigator for incidents and debugging. Uses hypothesis-driven methodology and systematic triage. Can query Axiom observability when available. Use for incident response, root cause analysis, production debugging, or log investigation.
github-pr-review
fvadicamo
Handles PR review comments and feedback resolution. Use when user wants to resolve PR comments, handle review feedback, fix review comments, address PR review, check review status, respond to reviewer, verify PR readiness, review PR comments, analyze review feedback, evaluate PR comments, assess review suggestions, or triage PR comments. Fetches comments via GitHub CLI, classifies by severity, applies fixes with user confirmation, commits with proper format, replies to threads.
sf-testing
Jaganpro
Comprehensive Salesforce testing skill with test execution, code coverage analysis, and agentic test-fix loops. Run Apex tests, analyze coverage, generate test patterns, and automatically fix failing tests with 120-point scoring.
codebase-memory
ToanMobile
Use the codebase knowledge graph for structural code queries, AST navigation, and blast radius analysis. Dùng khi khám phá code, tìm symbol, đọc implementation, trace call/data flow, ai gọi hàm này (inbound caller), hàm này gọi gì (outbound), phân tích blast radius, dead code, refactor candidates, Cypher query, hoặc xem kiến trúc dự án. Tự động fallback về Read/Grep khi thiếu node hoặc index stale.
fixbugs
ToanMobile
Quy trình chuẩn đoán, triage sự cố Crashlytics/ANR và sửa lỗi tuân thủ Paired Executable Oracle (RED -> GREEN). Kích hoạt khi gặp bug, crash log, production stack trace, ANR traces.txt, native crash, memory leak OOM, hoặc cần sửa lỗi phẫu thuật (surgical fix) không gây hồi quy.
deploy
ToanMobile
"Quy trình build, đóng gói APK/AAB, kiểm tra signing và chuẩn bị release cho dự án Android/Gradle. Không dùng cho dự án không phải Android."
deep-module-design
ToanMobile
Dùng khi thiết kế hoặc đánh giá interface/seam, testability hay abstraction của một module. Bỏ qua khi chỉ cần tìm caller/blast radius hoặc sửa local không đổi contract.
sf-ai-agentforce-observability
Jaganpro
Extract and analyze Agentforce session tracing data from Salesforce Data 360. Supports high-volume extraction (1-10M records/day), Polars-based analysis, and debugging workflows for agent sessions.
sf-debug
Jaganpro
Salesforce debugging and troubleshooting skill with log analysis, governor limit detection, and agentic fix suggestions. Parse debug logs, identify performance bottlenecks, analyze stack traces, and automatically suggest fixes.
android-real-device-qa
ToanMobile
Bộ công cụ kiểm thử chất lượng trên thiết bị Android thật hoặc máy ảo (emulator). Tự động kích hoạt khi kiểm thử trên Android, khi kết nối adb, đo FPS bề mặt (Jank/Stutter), trích xuất UI hierarchy XML để audit tọa độ/touch target, chẩn đoán ANR/crash logcat, quét method count DEX ngăn tràn 64K, và xuất báo cáo HTML kèm ảnh chụp thực tế. Tuyệt đối không tạo kết quả PASS ảo khi không có thiết bị thật online.
create-task
ToanMobile
"Tự động tạo task chuẩn Super-Board (Problem, Context, Fix, AC, Risk) và đẩy trực tiếp lên cột Backlog của GitHub Project Board #1 (https://github.com/users/ToanMobile/projects/1/views/2). Kích hoạt khi gõ /createTask, /create-task, hoặc khi người dùng yêu cầu 'tạo task', 'đẩy lên backlog'."
testing-patterns
groupzer0
TDD workflow and test strategy patterns including test pyramid, coverage strategies, mocking approaches, and anti-patterns. Load when writing tests, designing test strategies, or reviewing test coverage.
spec-driven-development
ToanMobile
Dùng khi feature/refactor/risky flow cần spec trước code, yêu cầu mơ hồ, hoặc thay đổi chạm từ 2 module, 3 file hay hơn 200 LOC net diff. Bỏ qua typo, config không đổi behavior và bug local đã rõ root cause.
code-review-standards
groupzer0
Code review checklist, severity definitions, and document templates. Load when performing code reviews or defining review criteria.
test-driven-development
Megatron-HQ
Use when implementing any feature or bugfix, before writing implementation code
release-procedures
groupzer0
Version management, release verification, and deployment procedures for software releases. Includes semver guidance, version consistency checks, and platform-specific constraints.
git-workflow
ScientiaCapital
"Git workflow patterns - conventional commits, branch naming, PR templates, merge strategies. Use when: commit, PR, branch naming, git workflow, conventional commits, semantic versioning."