
idahunt is a framework to analyze binaries with IDA Pro and hunt for things in IDA Pro

idahunt 是一个使用 IDA Pro 分析二进制文件并在 IDA Pro 中搜寻目标的框架。它是一个命令行工具,可递归分析给定文件夹中的所有可执行文件。它在后台执行 IDA,因此你无需手动打开每个文件。它支持执行外部 IDA Python 脚本。
有用的示例包括(非详尽列表):
IDA Python 脚本的能力是无限的。你可以导入任何现有的 IDA Python 脚本,或编写自己的脚本。一些示例:
idahunt.py:分析可执行文件的主工具filters/:包含基础过滤器,用于决定输入目录中的哪些文件需要用 IDA 分析
filters/default.py:默认的基础过滤器,不过滤任何内容,且默认使用filters/ciscoasa.py:用于分析 Cisco ASA 防火墙镜像filters/hpilo.py:用于分析 HP iLO 镜像filters/names.py:基于名称、名称长度或扩展名的基础过滤器script_template.py:包含一个 hello world IDA Python 脚本C:\idahunt> C:\Python37-x64\python.exe .\idahunt.py -h
usage: idahunt.py [-h] [--inputdir INPUTDIR] [--analyse] [--open]
[--ida-args IDA_ARGS] [--scripts SCRIPTS [SCRIPTS ...]]
[--filter FILTER] [--cleanup] [--temp-cleanup] [--verbose]
[--max-ida MAX_IDA] [--list-only] [--version IDA_VERSION]
optional arguments:
-h, --help show this help message and exit
--inputdir INPUTDIR Input folder to search for files
--analyse, --analyze analyse all files i.e. create .idb for all of them
--open open all files into IDA (debug only)
--ida-args IDA_ARGS Additional arguments to pass to IDA (e.g.
-p<processor> -i<entry_point> -b<load_addr>)
--scripts SCRIPTS [SCRIPTS ...]
List of IDA Python scripts to execute in this order
--filter FILTER External python script with optional arguments
defining a filter for the names of the files to
analyse. See filters/names.py for example
--cleanup Cleanup i.e. remove .asm files that we don't need
--temp-cleanup Cleanup temporary database files i.e. remove .id0,
.id1, .id2, .nam, .dmp files if IDA Pro crashed and
did not delete them
--verbose be more verbose to debug script
--max-ida MAX_IDA Maximum number of instances of IDA to run at a time
(default: 10)
--list-only List only what files would be handled without
executing IDA
--version IDA_VERSION
Override IDA version (e.g. "7.5"). This is used to
find the path of IDA on Windows.
你可以将 --list-only 与任何命令行结合使用,仅列出工具将执行的操作,而不会真正执行。
C:\idahunt>idahunt.py --inputdir C:\re --analyse --filter "filters\names.py -a 32 -v" --list-only
[idahunt] Simulating only...
[idahunt] ANALYSING FILES
[idahunt] Analysing C:\re\cves\cve-2014-4076.dll
[idahunt] Analysing C:\re\cves\cve-2014-4076.exe
[idahunt] Analysing C:\re\DownloadExecute.exe
[idahunt] Analysing C:\re\ReverseShell.exe
这里我们开始初始分析。它会在几秒后完成:
C:\idahunt>idahunt.py --inputdir C:\re --analyse --filter "filters\names.py -a 32 -v"
[idahunt] ANALYSING FILES
[idahunt] Analysing C:\re\cves\cve-2014-4076.dll
[idahunt] Analysing C:\re\cves\cve-2014-4076.exe
[idahunt] Analysing C:\re\DownloadExecute.exe
[idahunt] Analysing C:\re\ReverseShell.exe
[idahunt] Waiting on remaining 4 IDA instances
这里我们清理初始分析生成的临时 .asm 文件:
C:\idahunt>idahunt.py --inputdir C:\re --cleanup
[idahunt] Deleting C:\re\cves\cve-2014-4076.asm
[idahunt] Deleting C:\re\DownloadExecute.asm
[idahunt] Deleting C:\re\ReverseShell.asm
我们可以看到生成的 .idb 文件,以及一些包含 IDA Pro 输出窗口内容的 .log 文件。
C:\idahunt>tree /f C:\re
Folder PATH listing
Volume serial number is XXXX-XXXX
C:\RE
│ DownloadExecute.exe
│ DownloadExecute.idb
│ DownloadExecute.log
│ ReverseShell.exe
│ ReverseShell.idb
│ ReverseShell.log
│
└───cves
cve-2014-4076.dll
cve-2014-4076.exe
cve-2014-4076.idb
cve-2014-4076.log
这里我们执行一个基础的 IDA Python 脚本,它会在 IDA Pro 输出窗口中打印 [script_template] I execute in IDA, yay!。
C:\idahunt>idahunt.py --inputdir C:\re --filter "filters\names.py -a 32 -v" --scripts C:\idahunt\script_template.py
[idahunt] EXECUTE SCRIPTS
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\cves\cve-2014-4076.dll
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\cves\cve-2014-4076.exe
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\DownloadExecute.exe
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\ReverseShell.exe
[idahunt] Waiting on remaining 4 IDA instances
由于它被保存在 .log 文件中,我们可以确认其执行成功:
Autoanalysis subsystem has been initialized.
Database for file 'ReverseShell.exe' has been loaded.
Compiling file 'C:\Program Files (x86)\IDA 6.95\idc\ida.idc'...
Executing function 'main'...
[script_template] I execute in IDA, yay!
自这个 PR 起,idahunt 已与 diaphora 完美集成,用于进行二进制差异比较。
你需要一个包含同一文件不同版本的文件夹层级结构,例如:
C:\> tree C:\tests\ /F
C:\tests
├───patch
│ tm.sys
│
└───vuln
tm.sys
如果尚未完成,你需要先进行初始 IDA 分析以创建 IDB 文件。
C:\idahunt> python idahunt.py --inputdir C:\tests\ --analyse --verbose
[idahunt] IDA32 = C:\Program Files\IDA Core 8.1\ida.exe
[idahunt] IDA64 = C:\Program Files\IDA Core 8.1\ida64.exe
[idahunt] ANALYSING FILES
[idahunt] Analysing C:\tests\patch\tm.sys
[idahunt] C:\Program Files\IDA Core 8.1\ida64.exe -B -oC:\tests\patch\tm.i64 -LC:\tests\patch\tm.log C:\tests\patch\tm.sys
[idahunt] Analysing C:\tests\vuln\tm.sys
[idahunt] C:\Program Files\IDA Core 8.1\ida64.exe -B -oC:\tests\vuln\tm.i64 -LC:\tests\vuln\tm.log C:\tests\vuln\tm.sys
[idahunt] Executed IDA 2/2 times IDA instances
[idahunt] Took 0:00:15.03 to execute this
C:\> tree C:\tests\ /F
C:\tests
├───patch
│ tm.i64
│ tm.log
│ tm.sys
│
└───vuln
tm.i64
tm.log
tm.sys
这里使用 diaphora 对每个文件进行 diff 导出(创建 <filename>.sqlite sqlite3 数据库),然后进行版本间的 diff(创建 <filename>.diaphora sqlite3 数据库)。
C:\idahunt> python idahunt.py --diaphora-path C:\diaphora --inputdir C:\tests --diff --filename tm.sys --verbose
[idahunt] IDA32 = C:\Program Files\IDA Core 8.1\ida.exe
[idahunt] IDA64 = C:\Program Files\IDA Core 8.1\ida64.exe
[idahunt] EXECUTE DIFF-EXPORT
[idahunt] Executing script C:\diaphora\diaphora_ida.py for C:\tests\patch\tm.sys
[idahunt] C:\Program Files\IDA Core 8.1\ida64.exe -A -SC:\diaphora\diaphora_ida.py -LC:\tests\patch\tm.log C:\tests\patch\tm.i64
[idahunt] Environment variables:
[idahunt] DIAPHORA_AUTO2=1
[idahunt] DIAPHORA_EXPORT_FILE=tm.sqlite
[idahunt] Executing script C:\diaphora\diaphora_ida.py for C:\tests\vuln\tm.sys
[idahunt] C:\Program Files\IDA Core 8.1\ida64.exe -A -SC:\diaphora\diaphora_ida.py -LC:\tests\vuln\tm.log C:\tests\vuln\tm.i64
[idahunt] Environment variables:
[idahunt] DIAPHORA_AUTO2=1
[idahunt] DIAPHORA_EXPORT_FILE=tm.sqlite
[idahunt] Executed IDA 2/2 times IDA instances
[idahunt] EXECUTE DIFF
[idahunt] Diffing patch vs vuln
[idahunt] C:\Program Files\Python39\python.exe C:\diaphora\diaphora.py C:\tests\patch\tm.sqlite C:\tests\vuln\tm.sqlite -o C:\tests\vuln\patch_vs_vuln\tm.sys.diaphora
[diaphora][Wed Dec 14 10:33:49 2022] Diffing...es
[diaphora][Wed Dec 14 10:33:49 2022] Callgraphs from both programs differ in 0.706714%
[diaphora][Wed Dec 14 10:33:49 2022] Finding best matches...
[diaphora][Wed Dec 14 10:33:49 2022] Finding with heuristic 'Perfect match, same name'
[diaphora][Wed Dec 14 10:33:50 2022] All functions matched in at least one database, finishing.
[diaphora][Wed Dec 14 10:33:50 2022] Finding partial matches
[diaphora][Wed Dec 14 10:33:50 2022] All functions matched in at least one database, finishing.
[diaphora][Wed Dec 14 10:33:50 2022] Finding with heuristic 'Small names difference'
[diaphora][Wed Dec 14 10:33:50 2022] Finding with heuristic 'Call address sequence'
[diaphora][Wed Dec 14 10:33:50 2022] Finding with heuristic 'Call address sequence'
[diaphora][Wed Dec 14 10:33:50 2022] Finding unmatched functions
[diaphora][Wed Dec 14 10:33:50 2022] Done. Took 1.110000000000582 seconds.
[diaphora][Wed Dec 14 10:33:50 2022] Diffing results saved in file 'C:\tests\vuln\patch_vs_vuln\tm.sys.diaphora'.
[idahunt] Executed Python 1/1 times
[idahunt] Took 0:00:30.07 to execute this
C:\> tree C:\tests\ /F
C:\tests
├───patch
│ tm.i64
│ tm.log
│ tm.sqlite
│ tm.sys
│
└───vuln
│ tm.i64
│ tm.log
│ tm.sqlite
│ tm.sys
│
└───patch_vs_vuln
tm.sys.diaphora
tm.sys.txt
如上所示,它还会在 <filename>.diaphora 文件旁边创建一个 <filename>.txt 文件,其中包含最佳匹配列表:
partial,00000,1c0002708,WPP_SF_DDq,1c0002708,WPP_SF_DDq,0.950,1,1,Perfect match, same name
partial,00001,1c0002770,WPP_SF_Dq,1c0002770,WPP_SF_Dq,0.940,1,1,Perfect match, same name
partial,00002,1c00027c8,WPP_SF_qq_guid_D,1c00027c8,WPP_SF_qq_guid_D,0.960,1,1,Perfect match, same name
partial,00003,1c0002844,WPP_SF_qqi,1c0002844,WPP_SF_qqi,0.950,1,1,Perfect match, same name
partial,00004,1c00028a8,WPP_SF_qqii,1c00028a8,WPP_SF_qqii,0.960,1,1,Perfect match, same name
partial,00005,1c0015500,TmRecoverResourceManagerExt,1c0015500,TmRecoverResourceManagerExt,0.860,37,36,Perfect match, same name
partial,00006,1c001a610,TmpHeuristicAbortTransaction,1c001a640,TmpHeuristicAbortTransaction,0.986,6,6,Perfect match, same name
partial,00007,1c001a6c0,TmpHeuristicAbortTransactionAfterCheckpoint,1c001a6f0,TmpHeuristicAbortTransactionAfterCheckpoint,0.992,11,11,Perfect match, same name
partial,00008,1c001ad58,TmpIsClusteredTransactionManager,1c001ad88,TmpIsClusteredTransactionManager,0.994,15,15,Perfect match, same name
partial,00009,1c001b0c0,TmpMigrateEnlistments,1c001b0f0,TmpMigrateEnlistments,0.980,10,10,Perfect match, same name
事实证明,仓库中提供的 tm.sys 文件与 CVE-2018-8611 相关,因此让我们分析已修补的函数 TmRecoverResourceManagerExt():
C:\idahunt> python idahunt.py --diaphora-path C:\diaphora\ --inputdir C:\tests\ --html --filename tm.sys --funcname TmRecoverResourceManagerExt --verbose
[idahunt] IDA32 = C:\Program Files\IDA Core 8.1\ida.exe
[idahunt] IDA64 = C:\Program Files\IDA Core 8.1\ida64.exe
[idahunt] EXECUTE GENERATE HTML
C:\tests\patch\tm.sqlite C:\tests\vuln\tm.sqlite
C:\tests\vuln\patch_vs_vuln\tm.sys\TmRecoverResourceManagerExt_asm.html C:\tests\vuln\patch_vs_vuln\tm.sys\TmRecoverResourceManagerExt_pseudo.html
[idahunt] Showing patch vs vuln for TmRecoverResourceManagerExt
[idahunt] C:\Program Files\IDA Core 8.1\ida64.exe -A -SC:\diaphora\diaphora_ida.py -LC:\tests\patch\tm.log C:\tests\patch\tm.i64
[idahunt] Environment variables:
[idahunt] DIAPHORA_AUTO4=1
[idahunt] DIAPHORA_DB1=C:\tests\patch\tm.sqlite
[idahunt] DIAPHORA_DB2=C:\tests\vuln\tm.sqlite
[idahunt] DIAPHORA_DIFF=C:\tests\vuln\patch_vs_vuln\tm.sys.diaphora
[idahunt] DIAPHORA_EA1=1c0015500
[idahunt] DIAPHORA_EA2=1c0015500
[idahunt] DIAPHORA_HTML_ASM=C:\tests\vuln\patch_vs_vuln\tm.sys\TmRecoverResourceManagerExt_asm.html
[idahunt] DIAPHORA_HTML_PSEUDO=C:\tests\vuln\patch_vs_vuln\tm.sys\TmRecoverResourceManagerExt_pseudo.html
[idahunt] Executed IDA 1/1 times IDA instances
[idahunt] Took 0:00:10.03 to execute this
现在我们已生成该函数的汇编代码和反编译代码:
C:\idahunt> tree C:\tests\ /F
C:\tests
├───patch
│ tm.i64
│ ...
│
└───vuln
│ tm.i64
│ ...
│
└───patch_vs_vuln
│ tm.sys.diaphora
│ tm.sys.txt
│
└───tm.sys
TmRecoverResourceManagerExt_asm.html
TmRecoverResourceManagerExt_pseudo.html

我们可以让 idahunt 只分析文件名中包含指定模式的文件(下面的 -n Download):
C:\idahunt>idahunt.py --inputdir C:\re --filter "filters\names.py -a 32 -v -n Download" --scripts C:\idahunt\script_template.py --list-only
[idahunt] Simulating only...
[idahunt] EXECUTE SCRIPTS
[names] Skipping non-matching name Download in cve-2014-4076.dll
[names] Skipping non-matching name Download in cve-2014-4076.exe
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\DownloadExecute.exe
[names] Skipping non-matching name Download in ReverseShell.exe
我们也可以让 idahunt 只分析具有指定扩展名的文件(下面的 -e dll):
C:\idahunt>idahunt.py --inputdir C:\re --filter "filters\names.py -a 32 -v -e dll" --scripts C:\idahunt\script_template.py --list-only
[idahunt] Simulating only...
[idahunt] EXECUTE SCRIPTS
[idahunt] Executing script C:\idahunt\script_template.py for C:\re\cves\cve-2014-4076.dll
[names] Skipping non-matching extension .dll in cve-2014-4076.exe
[names] Skipping non-matching extension .dll in DownloadExecute.exe
[names] Skipping non-matching extension .dll in ReverseShell.exe
由于 IDA Pro 的架构,以及它包含两个不同的可执行文件 idaq.exe 和 idaq64.exe,分别用于分析 32 位和 64 位两种架构的二进制文件,因此需要提前知道目标架构。如果你想使用 HexRays 反编译器,这一点尤其重要。
idahunt 会自动检测 Windows PE 文件中的 i386、ia64 和 amd64 架构。如果你需要自动检测其他架构,可以创建 issue,或将其添加到 idahunt 并提交 PR。
如果你忘记提供要分析文件的架构,基础的 filters\names.py 将返回错误:
C:\idahunt>idahunt.py --inputdir C:\re --filter "filters\names.py -v -e dll" --scripts C:\idahunt\script_template.py --list-only
[idahunt] Simulating only...
[idahunt] EXECUTE SCRIPTS
[names] Unknown architecture: None. You need to specify it with -a
[names] Skipping non-matching extension .dll in cve-2014-4076.exe
[names] Skipping non-matching extension .dll in DownloadExecute.exe
[names] Skipping non-matching extension .dll in ReverseShell.exe
这里有用于分析 HP iLO 或 Cisco ASA 固件的过滤器。