使用方法
- 复制安装命令,在自己的智能体环境安装前先检查源码包。
- 从原始文档核对所需工具、凭据和支持的宿主。
- 说明输入材料与期望交付物,并对照来源检查输出。
npx skills add https://github.com/smartworldarafath/rea_skill --skill rea-performance-crash-triage开始前需了解
- 安装不包含模型订阅、API 额度或访问文件的权限。
- 仓库星数和安装量反映活跃情况,不等于独立质量或安全测试。
按源仓库内容呈现,保留标题、案例、代码、表格、链接以及原文引用的演示图片。
REA Crash Forensics, Runtime Instrumentation & Performance Profiling Skill
Scope & Authorization
Covers all first-party crash dump analysis, live runtime debugging, Frida hooking, and CPU/memory performance optimization tasks (including Category 9 of the 500+ REA Prompt Vault and any custom debugging/profiling request).
Execution Pipeline
- Crash Dump & Exception Forensics:
- Use REA MCP tools:
inspect_recorded_crash,open_binary,resolve_containing_procedure,read_function_instructions, andinspect_address_context. - Decode exception codes (
0xC0000005Access Violation,SIGSEGV, heap corruption, stack overflow), unwind stack frames, map faulting instruction pointers (RIP/EIP/PC) to decompiled functions, and produce a root-cause fix.
- Dynamic Runtime Instrumentation (Frida / GDB / Rizin):
- Use REA MCP tools:
list_frida_devices,list_frida_processes,start_frida_session,instrument_with_frida,observe_native_calls,start_gdb_session,gdb_console,start_rizin_debug_session,rizin_debug_command,capture_process_scenario, andcorrelate_static_and_runtime. - Trace live API calls, argument values, return buffers, and UI event dispatch (
observe_native_ui,trace_native_ui_action).
- Performance Profiling & Bottleneck Elimination:
- Profile hot call paths using
flameoxflamegraphs and REA call graphs (get_call_graph). - Refactor O(N^2) loops, UI thread blocking calls, and heap contention into verified, zero-skeleton optimized implementations with benchmark tests.