← ClaudeAtlas

ctf-rev-toolkitlisted

Reverse engineering toolkit and workflow for CTF rev. Load for any rev challenge to pick the right approach by target type (native ELF/PE, .NET, JVM, Python pyc, WASM, packed/obfuscated), find the flag-gating check, and recover the required input, including angr symbolic execution for constraint-heavy binaries.
bilal-spotted/claude-code-ctf-rig · ★ 3 · AI & Automation · score 69
Install: claude install-skill bilal-spotted/claude-code-ctf-rig
# CTF Reverse Engineering Toolkit ## Triage (always first) - `file`, `strings -n 6` (flag or hints sometimes sit in plaintext). - `checksec`, detect packing (entropy, UPX signature). UPX -> `upx -d`. - Identify runtime/language; it dictates the tool. ## By target type - Native ELF/PE: Ghidra (decompile, rename as you learn) via ghidra MCP; radare2/gdb+pwndbg for dynamic. Watch for anti-debug (ptrace, timing). - .NET: ilspycmd or dnSpy; IL is high-level, often trivial once decompiled. - JVM .class/.jar: cfr or procyon decompiler. - Python .pyc: decompyle3 / pycdc; match bytecode version. - WASM: wasm2wat, read the text form. - Go/Rust: stripped and large; find main.main, use symbol recovery scripts, focus on the comparison logic not the runtime. ## Find the gate Locate where input is validated: strcmp/memcmp, a checksum loop, a transform-then-compare. Explain what it checks in NOTES.md before solving. ## Recover input - Direct compare -> read the target bytes, done. - Simple transform (xor/add/permute) -> invert it. - Many interlocking constraints on input bytes -> angr: ```python import angr, claripy p = angr.Project('./chall', auto_load_libs=False) # find = address of "Correct" path, avoid = "Wrong" path sm = p.factory.simulation_manager(p.factory.full_init_state()) sm.explore(find=FIND_ADDR, avoid=AVOID_ADDR) print(sm.found[0].posix.dumps(0)) # stdin that solves it ``` - Verify: run the binary with the recovered input, confirm success path. ## Discipline - If you