noorqureshi/security-skills/ctf-rev
Claude Code and agent skill by noorqureshi/security-skills
- community
Checking that the SKILL.md file is still in the repository.
CTF reversing speed-run — crack flag-check binaries fast: locate the check via strings/xrefs, bypass anti-debug and obfuscation, and solve constraint-checkers with z3 instead of manual algebra. Load when the handout is an ELF/PE/exe that asks for input or a key. Signals: "rev" category, "Enter the flag/key/password", correct/wrong messages, crackme, packed binary, anti-debug (ptrace, IsDebuggerPresent), keygen challenge.
Install
npx vskill@latest install noorqureshi/security-skills/ctf-revSKILL.md
CTF reverse engineering speed-run
When it applies
The handout is a binary that checks something — a flag, key, serial, or password — and you need the accepted input. This skill is the CTF workflow: find the check fast, defang whatever stops you from seeing it, and solve the constraints mechanically. For deep tool use defer to reverse-eng-binary-triage (first-pass recon), reverse-eng-radare2 / reverse-eng-ghidra / reverse-eng-ida (analysis), reverse-eng-deobfuscation (unpacking/devirtualization).
Why it works
A flag-checker is a pure function from input to accept/reject, and CTF ones are small. You never need to understand the whole binary — only the check — so the game is locating it (strings and xrefs beat linear reading), exposing it (anti-debug and packing are thin in CTF), and solving it (the constraints are almost always z3-shaped; manual algebra is the slow path).
Method
- Two-minute triage.
file,strings -n 6(+strings -elfor UTF-16),checksec, run it
with junk input. The strings that matter: "Correct!"/"Wrong", the flag format itself (sometimes just… present), odd alphabets, and packer markers (UPX!). Packed → upx -d first; anything fancier → reverse-eng-deobfuscation. .NET/Java/Go/Rust/Python-exe binaries route to their dedicated tooling (reverse-eng-dotnet, reverse-eng-go-rust, pyinstxtractor+uncompyle).
- Locate the check, don't read the binary. In r2/Ghidra/IDA: find the "Wrong" string, follow
the xref to its function, and read backwards from the branch that prints it. Everything between the input read and that branch is the check. For C/C++ this is usually one function.
- Dynamic shortcuts before static grinding.
ltrace ./chall— library calls likestrcmp(input, "..."),strlen, per-char compares
leak the answer or its length with zero analysis.
- Compare-per-character binaries → count crashes/coverage: flip one input byte at a time and
watch which comparison index changes, or use a debugger on the compare loop to read each expected byte as it's checked.
stracefor flag files, env vars, network the binary secretly consults.
- Defang anti-debug (it's shallow in CTF). Common tricks and one-line counters:
ptrace(PTRACE_TRACEME)→ patch the call, orcatch syscall ptracein gdb and force
return 0.
IsDebuggerPresent/PEB.BeingDebugged→ zero the flag in the debugger, or patch the
conditional jump that follows.
- Timing checks (
rdtsc) → patch or ignore — they only guard a branch. - Self-checksum / anti-tamper → make all your patches before the checksum computes, or patch
the comparison result, not the data. The universal move: find the check that branches on the anti-debug result and flip that one branch — don't fight each mechanism.
- Solve the check with z3 — the CTF superpower. When the check transforms input bytes and
compares against constants (xor/add/rotate/substitute chains, matrix math, per-position relations), transcribe the decompiled pseudocode into z3 almost verbatim: ``python from z3 import * flag = [BitVec(f'c{i}', 8) for i in range(32)] s = Solver() # constraints copied from the decompilation, e.g.: s.add((flag[i] ^ 0x37) + i & 0xff == target[i]) s.add(And(*[And(0x20 <= c, c < 0x7f) for c in flag])) # printable s.check(); m = s.model(); print(bytes(m[c].as_long() for c in flag)) ` Use BitVec (not Int) so overflow wraps like machine arithmetic. If the decompile is messy, feed the whole function to angr (angr with find=<success addr>, avoid=<fail addr>`) and let it symbolically execute to the answer.
- Know the recurring shapes. Flag = input satisfying equations (z3); keygen (invert the
serial transform, then generate any valid one); VM challenges (small custom bytecode — write a 30-line disassembler for the handlers, the flag falls out); wasm (wasm2c/wabt, then same game); mobile crackmes (mobile-apk-reverse).
Gotchas
- The first "Correct" string can be a decoy — verify your answer actually prints success on
the real binary; some challenges have fake success paths guarding the real check.
- Anti-debug that crashes you *is* the solve path — if debugging changes behavior, run under
the debugger anyway and note where; the flag transformation sometimes only happens under ptrace.
- BitVec vs Int — z3
Intdoesn't wrap; forgetting& 0xffsemantics makes solvers return
unsat on trivially solvable checks.
- Multiple valid flags — constraint checks can accept many inputs; the scoreboard wants the
one in flag format, so add flag{-prefix constraints when known.
- Don't statically invert what you can dynamically read — a
strcmpin ltrace beats an hour
of decompilation; try dynamic first, always.
- Stripped/optimized binaries mislead decompilers — check the disassembly when pseudocode
looks impossible (reverse-eng-binary-triage for triage discipline).
Verify success
The recovered input, entered into the original unmodified binary, produces the success path — and matches the event flag format. For keygens, two independently generated keys both validate.
References
Triage: reverse-eng-binary-triage; deep analysis: reverse-eng-ghidra, reverse-eng-ida, reverse-eng-radare2; unpacking/anti-analysis: reverse-eng-deobfuscation; z3 docs; angr. Triage via ctf-methodology.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/ctf/ctf-rev/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/noorqureshi/security-skills/ctf-rev)