reverse-skills-max: initial public release (restored, redacted, hardened)

reverse-skill-repo/scripts/pe_deep.py

@@ -0,0 +1,103 @@
+#!/usr/bin/env python3
+"""Deep PE static reverse without IDA: capstone linear+entry disasm, IAT thunk
+resolution, call-site mapping to imports, and payload/container carving.
+
+Usage: python pe_deep.py <path.exe> [--disasm-entry N] [--carve]
+"""
+from __future__ import annotations
+import argparse, sys
+import pefile
+from capstone import Cs, CS_ARCH_X86, CS_MODE_64, CS_MODE_32
+
+
+def rva_to_off(pe, rva):
+    return pe.get_offset_from_rva(rva)
+
+
+def build_iat_map(pe):
+    """Map IAT slot VA -> imported name (for call [rip+x] / call [iat] resolution)."""
+    m = {}
+    if not hasattr(pe, "DIRECTORY_ENTRY_IMPORT"):
+        return m
+    base = pe.OPTIONAL_HEADER.ImageBase
+    for entry in pe.DIRECTORY_ENTRY_IMPORT:
+        dll = entry.dll.decode("latin1", "replace")
+        for imp in entry.imports:
+            nm = imp.name.decode("latin1", "replace") if imp.name else f"{dll}#ord{imp.ordinal}"
+            if imp.address is not None:
+                m[imp.address] = f"{dll}!{nm}"          # absolute VA of the IAT slot user
+            # imp.address is the VA where the pointer is used; also map the thunk RVA
+    return m
+
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("path")
+    ap.add_argument("--disasm-entry", type=int, default=120, help="instructions from entrypoint")
+    ap.add_argument("--carve", action="store_true")
+    args = ap.parse_args()
+
+    pe = pefile.PE(args.path, fast_load=False)
+    is64 = hex(pe.OPTIONAL_HEADER.Magic) == "0x20b"
+    md = Cs(CS_ARCH_X86, CS_MODE_64 if is64 else CS_MODE_32)
+    md.detail = True
+    base = pe.OPTIONAL_HEADER.ImageBase
+    ep_rva = pe.OPTIONAL_HEADER.AddressOfEntryPoint
+    ep_va = base + ep_rva
+    print(f"# {args.path}")
+    print(f"imagebase={hex(base)} entry_va={hex(ep_va)} bits={'64' if is64 else '32'}")
+
+    # Build a name map for IAT slots keyed by the pointer VA the code dereferences.
+    # pefile: imp.address = VA of the IAT entry (where the resolved pointer lives).
+    iat = {}
+    if hasattr(pe, "DIRECTORY_ENTRY_IMPORT"):
+        for entry in pe.DIRECTORY_ENTRY_IMPORT:
+            dll = entry.dll.decode("latin1", "replace")
+            for imp in entry.imports:
+                nm = imp.name.decode("latin1", "replace") if imp.name else f"ord{imp.ordinal}"
+                if imp.address:
+                    iat[imp.address] = f"{dll}!{nm}"
+
+    # Disassemble from entrypoint.
+    ep_off = rva_to_off(pe, ep_rva)
+    code = pe.get_memory_mapped_image()[ep_rva: ep_rva + args.disasm_entry * 16]
+    print(f"\n== entrypoint disasm ({args.disasm_entry} insns) ==")
+    n = 0
+    for insn in md.disasm(code, ep_va):
+        annot = ""
+        # resolve call/jmp [rip+disp] against IAT
+        if insn.mnemonic in ("call", "jmp") and insn.operands:
+            op = insn.operands[-1]
+            try:
+                from capstone.x86 import X86_OP_MEM, X86_OP_IMM
+                if op.type == X86_OP_MEM and op.mem.base in (0,) and is64:
+                    # rip-relative: target = next_insn + disp
+                    tgt = insn.address + insn.size + op.mem.disp
+                    if tgt in iat:
+                        annot = f"  ; -> {iat[tgt]}"
+                elif op.type == X86_OP_MEM:
+                    if op.mem.disp in iat:
+                        annot = f"  ; -> {iat[op.mem.disp]}"
+                elif op.type == X86_OP_IMM and op.imm in iat:
+                    annot = f"  ; -> {iat[op.imm]}"
+            except Exception:
+                pass
+        print(f"  {insn.address:#0x}: {insn.mnemonic:<7}{insn.op_str}{annot}")
+        n += 1
+        if n >= args.disasm_entry:
+            break
+
+    if args.carve:
+        print("\n== carve scan ==")
+        rd = [s for s in pe.sections if s.Name.rstrip(b'\x00') == b'.rdata']
+        if rd:
+            data = rd[0].get_data()
+            print(f".rdata off={hex(rd[0].PointerToRawData)} size={len(data)}")
+            # leading dword table
+            import struct
+            vals = struct.unpack_from("<8I", data, 0)
+            print("leading dwords:", [hex(v) for v in vals])
+
+
+if __name__ == "__main__":
+    sys.exit(main())