# Detecting Dead Code > Identifies unreachable functions, unused variables, and abandoned code in binary programs. Use when optimizing binary size, analyzing code coverage, or investigating abandoned functionality. Use when this capability is needed. - Skill: `tomevault-io/detecting-dead-code` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/detecting-dead-code` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/detecting-dead-code/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/detecting-dead-code --- # Dead Code Detection ## Detection Workflow 1. **Build call graph**: Use `disasm` to build control flow, identify all functions, map function call relationships 2. **Identify entry points**: Find main/entry functions, locate exported functions, identify callback registrations, map interrupt handlers 3. **Trace reachability**: Perform forward reachability analysis from entry points, mark all reachable functions and blocks, identify unreachable code 4. **Analyze dead code**: Categorize dead code types, assess security implications, identify potential vulnerabilities, estimate size reduction ## Key Patterns - Unreachable functions: functions with no callers, functions only called from dead code, functions after infinite loops, functions after exit() calls - Unreachable blocks: basic blocks after return statements, code after unconditional jumps, blocks with no incoming edges, code guarded by always-false conditions - Unused data: global variables never referenced, string constants never used, data sections with no references, debug symbols in production builds - Conditional dead code: code in always-false branches, debug-only code in release builds, platform-specific code for other platforms, feature flags permanently disabled ## Output Format Report with: id, type, subtype, severity, confidence, location, dead_code_type, reason, callers, references, size_estimate, security_implications, potential_vulnerabilities, recommendation. ## Severity Guidelines - **MEDIUM**: Dead code containing security vulnerabilities - **LOW**: Dead code with no security impact ## See Also - `patterns.md` - Detailed detection patterns and exploitation scenarios - `examples.md` - Example analysis cases and code samples - `references.md` - CWE references and mitigation strategies --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/waiwai24) — claim your Tome and manage your conversions.