dwarf-expert
Analyzes DWARF debug information in compiled binaries. Use when inspecting .debug_* sections, DIE trees, or DW_TAG_/DW_AT_ entries with dwarfdump/llvm-dwarfdump or readelf, verifying debug info with llvm-dwarfdump --verify, answering DWARF standard questions, or writing code that parses DWARF (libdwarf, pyelftools, gimli).
- License 许可
- CC-BY-SA-4.0
- 最近更新
- 2026-11-06
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- dwarf-expert
- effort
- medium
- description
- Analyzes DWARF debug information in compiled binaries. Use when inspecting .debug_* sections, DIE trees, or DW_TAG_/DW_AT_ entries with dwarfdump/llvm-dwarfdump or readelf, verifying debug info with llvm-dwarfdump --verify, answering DWARF standard questions, or writing code that parses DWARF (libdwarf, pyelftools, gimli).
- allowed-tools
- Read Write Edit Bash Grep Glob WebSearch WebFetch
02 资源文件 2
- 参考文档agents/openai.yaml128 B
- 资产文件assets/trail-of-bits-mark.svg3.0 KB
03 SKILL.md 正文
DWARF Expert
Expertise for DWARF debug info: parsing and searching it, verifying its integrity, answering questions about the standard, and writing code that consumes it. Out of scope: runtime debugging (use gdb/lldb), reverse engineering beyond the DWARF sections (use Ghidra/IDA), and compiler-specific DWARF generation bugs.
Authoritative Sources
When precision matters, look standard details up instead of answering from memory:
- dwarfstd.org — the official specification. Web-search specific sections, e.g. "DWARF5 DW_TAG_subprogram attributes site:dwarfstd.org".
- LLVM —
llvm/lib/DebugInfo/DWARF/is a reliable reference implementation:DWARFDie.cpp(DIE and attribute access),DWARFUnit.cpp(compilation units),DWARFDebugLine.cpp(line tables),DWARFVerifier.cpp(validation). - libdwarf — the reference C implementation at github.com/davea42/libdwarf-code.
Parsing and Searching with dwarfdump
Prefer dwarfdump over readelf for DWARF-specific work. Two implementations
exist — libdwarf's dwarfdump and LLVM's llvm-dwarfdump — with different
options, and a bare dwarfdump command may be either: check dwarfdump --version
first. The options below are LLVM's.
On macOS, linked Mach-O executables do not carry DWARF: it stays in the .o
files until dsymutil collects it into a .dSYM bundle. Point dwarfdump at
the dSYM (or the object files), not the executable. pyelftools is ELF-only —
for Mach-O scripted work, stay with the LLVM tools.
--all: dump every DWARF section;--debug-info,--debug-line, etc. dump one--show-children [--recurse-depth=<n>]: include child DIEs when printing selected entries — parameters, locals, and struct members are children of function and type DIEs--show-parents [--parent-recurse-depth=<n>]: include parent DIEs--show-form: print attribute form types, for when encoding details matter--find=<name>: exact-name lookup via the accelerator tables — fast but not exhaustive; fall back to--namewhen it misses--name=<pattern> [--ignore-case] [--regex]: exhaustive DIE-name search--lookup=<address>: find the DIE covering an address--verbose: print low-level encoding detail
Searching DIEs
Escalate through these strategies as the query grows more complex:
- Name or address match:
--find, then--name;--lookupfor addresses. - Attribute or type queries (e.g. all parameters of type
float *): dump and filter.grep -Bpulls in the header line carrying each DIE's offset:llvm-dwarfdump file | grep -B 5 "float \*" | grep DW_TAG_formal_parameter, then print each DIE at its offset with--debug-info=<offset> --show-children(--lookuptakes a program address, not a DIE offset). - Multi-attribute or structural queries: when grep pipelines turn brittle,
write a Python script using
pyelftoolsinstead.
Verifying DWARF Integrity
llvm-dwarfdump --verify <binary>: structural checks (unit chains, DIE relationships, address ranges).--error-display=<quiet|summary|details|full>controls detail;--verify-json=<path>writes a machine-readable error summary;--quietfor exit-code-only checks.llvm-dwarfdump --statistics <binary>: debug-info quality metrics as JSON — compare across compiler versions or optimization levels to catch regressions.
Verify after producing DWARF (compilers, binary rewriters), when a debugger misbehaves on a binary, and when developing DWARF tooling against known-good files.
When a current-generation compiler emitted an old DWARF version, the build
explicitly passed -gdwarf-N — modern gcc and clang default to v4/v5, so check
the build system rather than assuming a toolchain default. GCC embeds its flags
in DW_AT_producer, so the pin is often readable right there; clang's producer
string carries no flags. Old versions remain common in the wild and read the
same way apart from surface forms: in v2 output, member offsets appear as
location expressions (DW_OP_plus_uconst) and linkage names as
DW_AT_MIPS_linkage_name.
readelf
For general ELF structure, or when dwarfdump is unavailable:
--debug-dump=<section>: dump a DWARF section (info,line, ...)--dwarf-depth=<n>/--dwarf-start=<n>: limit DIE depth / start offset
Writing Code That Parses DWARF
Prefer an existing library over parsing by hand:
| Library | Language | Notes |
|---|---|---|
libdwarf |
C/C++ | github.com/davea42/libdwarf-code — low-level; used to implement dwarfdump |
pyelftools |
Python | github.com/eliben/pyelftools — also parses ELF in general |
gimli |
Rust | github.com/gimli-rs/gimli — pair with object to load container files |
debug/dwarf |
Go | standard library |
LibObjectFile |
.NET | github.com/xoofx/LibObjectFile — also handles ELF/PE object files |
Default to Python with pyelftools for one-off scripts unless the task dictates
otherwise.
DWARF-specific pitfalls to handle — and to check for when reviewing DWARF code:
- Attributes are optional: a DIE may omit
DW_AT_name,DW_AT_type, ranges, etc. - Attribute indirection: a DIE's attributes may live on the DIE referenced by its
DW_AT_abstract_origin(inlined instances) orDW_AT_specification(out-of-line definitions) — resolve the chain before concluding data is absent. - Type chains: qualifiers and modifiers (
DW_TAG_const_type,DW_TAG_pointer_type, ...) wrap the underlying type; walkDW_AT_typelinks to reach the base type.
04 同仓库更多 Skills
zeroize-audit
Detects missing zeroization of sensitive data in source code and identifies zeroization removed by compiler optimizations, with assembly-level analysis, and control-flow verification. Use for auditing C/C++/Rust code handling secrets, keys, passwords, or other sensitive data.
yara-rule-authoring
Guides authoring of high-quality YARA-X detection rules for malware identification. Use when writing, reviewing, or optimizing YARA rules. Covers naming conventions, string selection, performance optimization, migration from legacy YARA, and false positive reduction. Triggers on: YARA, YARA-X, malware detection, threat hunting, IOC, signature, crx module, dex module.
writing-lean-proofs
Writes and reviews structured Lean 4 proofs and designs Lean libraries following Mathlib conventions. Use when proving theorems in Lean, formalizing mathematics or specifications in Lean 4, defining new types or definitions in a Lean library, reviewing Lean proofs for readability and maintainability, refactoring long tactic proofs into lemmas, filling in sorry placeholders in a Lean development, setting up CI or linters for a Lean project, diagnosing slow proofs or maxHeartbeats timeouts, or writing custom tactics, macros, or linters.
vulnerability-triage-brocards
This skill should be used when the user asks to "triage a vulnerability report", "assess a CVE", "evaluate a bug bounty submission", "decide if a finding is valid", "review a security finding", "dismiss a vulnerability", "should we fix this CVE", "prioritize a vulnerability report", or needs to determine whether an incoming vulnerability report warrants investigation. Applies 7 brocards (rules of thumb) to systematically accept, dismiss, or request more information on vulnerability reports, or needs to filter raw findings from agentic vulnerability discovery pipelines before human review.
variant-analysis
Hunts for the other instances of a bug already found — the variants of one root cause across a codebase. Use immediately after a vulnerability, logic bug, or bad pattern turns up in a specific file and the question becomes where else it occurs, including the bare conversational form ("are there others like this?", "is this the same bug?"). Also for generalizing one known instance into a CodeQL or Semgrep query for its whole pattern family, and for triaging a set of look-alike candidates against a known root cause. Not for initial discovery with no bug in hand.
vector-forge
Mutation-driven test vector generation. Finds implementations of a cryptographic algorithm or protocol, runs mutation testing to identify escaped mutants, then generates new test vectors that deliberately exercise the uncovered code paths. Compares before/after mutation kill rates to prove vector effectiveness. Use when generating cryptographic test vectors, measuring Wycheproof coverage gaps, finding escaped mutants via mutation testing, creating cross-implementation test suites, or improving test vector coverage for crypto primitives.