detecting-memory-leaks
通过识别错误路径中未释放的内存分配和缺失的清理来检测内存泄漏漏洞。在分析长时间运行的进程、资源管理或调查内存耗尽问题时使用。
📖 中文导读:通过识别错误路径中未释放的内存分配和缺失的清理来检测内存泄漏漏洞。在分析长时间运行的进程、资源管理或调查内存耗尽问题时使用。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。
name: detecting-memory-leaks description: Detects memory leak vulnerabilities by identifying unfreed memory allocations and missing cleanup in error paths. Use when analyzing long-running processes, resource management, or investigating memory exhaustion issues.
Memory Leak Detection
Detection Workflow
- Identify allocation points: Find all malloc/calloc/realloc calls, new/delete pairs (C++), map allocation locations
- Trace allocation lifecycle: Use
xrefs_toto trace allocations, follow pointers through code, identify all free() calls - Check cleanup paths: Verify every allocation has a corresponding free, check all error handling paths, review exception handling, assess callback cleanup
- Assess impact: How much memory is leaked? How often does it leak? What's the cumulative impact? Can it cause DoS?
Key Patterns
- Missing free: malloc() without corresponding free(), calloc() without free(), new without delete (C++)
- Exception path leaks: allocations before exception-prone operations, missing cleanup in error handling, leaks in early returns
- Reference cycles: circular references in data structures, reference counting bugs, shared pointer cycles
- Long-lived allocations: allocations that never get freed, caches without size limits, accumulating data structures
Output Format
Report with: id, type, subtype, severity, confidence, location, allocation (function, address, size), leak path, leak scenario, leak size per call, frequency, exploitability, impact, mitigation.
Severity Guidelines
- HIGH: Large leaks in frequently called functions
- MEDIUM: Small leaks or infrequent allocations
- LOW: Minor leaks with limited impact
See Also
patterns.md- Detailed detection patterns and exploitation scenariosexamples.md- Example analysis cases and code samplesreferences.md- CWE references and mitigation strategies