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

reverse-skill-repo/scripts/pe_behavior.py

@@ -0,0 +1,82 @@
+#!/usr/bin/env python3
+"""Behavior-focused static reverse: resolve the real main from the CRT stub,
+map every import call-site in .text to its API, dump resource directory + carve
+embedded PE/streams. No IDA. x64 only for call-site RIP resolution.
+
+Usage: python pe_behavior.py <path.exe>
+"""
+from __future__ import annotations
+import sys, struct
+import pefile
+from capstone import Cs, CS_ARCH_X86, CS_MODE_64
+from capstone.x86 import X86_OP_MEM, X86_OP_IMM, X86_REG_RIP
+
+
+def main():
+    path = sys.argv[1]
+    pe = pefile.PE(path, fast_load=False)
+    base = pe.OPTIONAL_HEADER.ImageBase
+    md = Cs(CS_ARCH_X86, CS_MODE_64); md.detail = True
+
+    # IAT slot VA -> name
+    iat = {}
+    for entry in getattr(pe, "DIRECTORY_ENTRY_IMPORT", []) or []:
+        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.split('.')[0]}!{nm}"
+
+    text = [s for s in pe.sections if s.Name.rstrip(b'\x00')==b'.text'][0]
+    code = text.get_data()
+    text_va = base + text.VirtualAddress
+
+    # Count import call-sites across whole .text (linear sweep).
+    from collections import Counter
+    calls = Counter()
+    callsites = {}   # name -> [addrs]
+    for insn in md.disasm(code, text_va):
+        if insn.mnemonic in ("call","jmp") and insn.operands:
+            op = insn.operands[-1]
+            tgt = None
+            if op.type == X86_OP_MEM and op.mem.base in (0, X86_REG_RIP):
+                tgt = insn.address + insn.size + op.mem.disp
+            elif op.type == X86_OP_IMM:
+                tgt = op.imm
+            if tgt in iat:
+                nm = iat[tgt]; calls[nm]+=1
+                callsites.setdefault(nm, []).append(insn.address)
+
+    print(f"# {path}")
+    print(f"imagebase={hex(base)} .text_va={hex(text_va)} size={len(code)}")
+    print("\n== import call-site frequency (behavior surface) ==")
+    for nm,c in calls.most_common():
+        addrs = ",".join(hex(a) for a in callsites[nm][:4])
+        print(f"  {c:3}x  {nm:<32} @ {addrs}")
+
+    # Resource directory dump + carve
+    print("\n== resources ==")
+    if hasattr(pe,"DIRECTORY_ENTRY_RESOURCE"):
+        for t in pe.DIRECTORY_ENTRY_RESOURCE.entries:
+            tn = pefile.RESOURCE_TYPE.get(t.id, t.id)
+            for r in getattr(t,"directory",{}).entries if hasattr(t,"directory") else []:
+                for l in getattr(r,"directory",{}).entries if hasattr(r,"directory") else []:
+                    d = l.data.struct
+                    print(f"  type={tn} id={r.id} size={d.Size} rva={hex(d.OffsetToData)}")
+
+    # Whole-file embedded PE scan (extractable child?)
+    b = open(path,"rb").read()
+    print("\n== embedded PE scan ==")
+    cnt=0
+    for i in range(0, len(b)-0x100):
+        if b[i:i+2]==b'MZ':
+            e = i + int.from_bytes(b[i+0x3c:i+0x40],'little')
+            if 0<e<len(b)-4 and b[e:e+4]==b'PE\x00\x00':
+                mach = int.from_bytes(b[e+4:e+6],'little')
+                if mach in (0x8664,0x14c):
+                    cnt+=1
+                    if cnt<=8: print(f"  child PE @ file+{hex(i)} machine={hex(mach)}")
+    print(f"  total embedded PE headers: {cnt}")
+
+
+if __name__ == "__main__":
+    main()