- name
SKILL.md
The skill file stays in the publisher's repository.
Paste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add index-login/MobileRE-Skill --skill dex-mem-dump -a cursor -yPaste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add index-login/MobileRE-Skill --skill dex-mem-dump -a claude-code -yPaste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add index-login/MobileRE-Skill --skill dex-mem-dump -a codex -yPaste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add index-login/MobileRE-Skill --skill dex-mem-dump -a grok -ySend the copied message in a Grok Bot chat. The Bot fetches this SKILL.md from GitHub and adds it to the skills all your Bots share.
Install this skill: https://github.com/index-login/MobileRE-Skill/tree/HEAD/.kilo/skill/dex-mem-dumpFrom the publisher
从运行中进程的内存里抠 DEX。两个 arm64 工具,都不走 ptrace:
二者都不用
ptrace。目标即使已用PTRACE_TRACEME占住 ptrace 槽(TracerPid != 0)也挡不住;能挡住的是真正基于 ptrace 的 dumper。二进制与源码在本 skill 目录;
mem-dex-dumper为本仓库自研。
边界:本卡是被动取快照,tools/unpack.py 是主动触发回填——后者靠 loadClass 让壳把方法体写回 DEX,所以能补出前者原理上拿不到的内容;代价是必须过 Frida 检测。对"按方法执行粒度回填"的高级壳,两者都拿不到,转 native。升级路径:分诊 → 本卡(快照)→ 方法体齐全则结束 → 骨架则 tools/unpack.py(回填)→ 仍骨架 = 高级壳。
分诊与升级细节见 references/unpacking.md ›「Step 1 root 内存 dump(非 Frida,优先)」。
adb devices 有设备;多设备所有命令加 -s <serial>。adb shell su -c 'id' 返回 uid=0。adb push .kilo/skill/dex-mem-dump/panda-dex-dumper /data/local/tmp/
adb shell chmod +x /data/local/tmp/panda-dex-dumper
mem-dex-dumper 同理,只换文件名。
adb shell pidof <包名>
adb shell monkey -p <包名> -c android.intent.category.LAUNCHER 1# 主选:panda(SIGSTOP 冻结快照)
adb shell "cd /data/local/tmp && ./panda-dex-dumper -p <pid>"
# 备选:mem(不冻结;-b mem 默认,-b vmreadv 换后端)
adb shell "su -c 'cd /data/local/tmp && ./mem-dex-dumper -p <pid> -o /data/local/tmp/memdump'"
adb pull /data/local/tmp/panda/ ./
adb pull /data/local/tmp/memdump/ ./
adb shell "su -c 'rm -rf /data/local/tmp/panda /data/local/tmp/memdump /data/local/tmp/panda-dex-dumper /data/local/tmp/mem-dex-dumper'"
更细的排障(PR_SET_DUMPABLE、hidepid、多进程)见 references/unpacking.md ›「Step 1 root 内存 dump(非 Frida,优先)」。
<包名>/dex/(或用户指定目录),保留工具原始命名。python3 tools/dex_rebuilder.py --fix-checksum <dump.dex> -o fixed.dex,否则 jadx 报 Bad dex file checksum。REPORT.md 记:工具、后端、命中地址 → maps 归属、md5 交叉校验。扫描/校验细节与重建步骤见 references/unpacking.md ›「工具机制与重建」。
运行时 DEX 内存 dump / 脱壳(root 直读 /proc/<pid>/mem,不注入、不 ptrace):常规解包拿不到、抽取壳未回填、或目标查 TracerPid 时用。流程:推工具 → 跑 → 拉回 → 清理;分诊与排障见 references/unpacking.md。
---name: dex-mem-dumpdescription: 运行时 DEX 内存 dump / 脱壳(root 直读 /proc/<pid>/mem,不注入、不 ptrace):常规解包拿不到、抽取壳未回填、或目标查 TracerPid 时用。流程:推工具 → 跑 → 拉回 →…---# .kilo/skill/dex-mem-dump/SKILL.mdFAQ
Download .kilo/skill/dex-mem-dump/SKILL.md from the publisher, or copy the path and open the repository. PlugMCP lists the skill and does not host the file.
The publisher keeps it at https://github.com/index-login/MobileRE-Skill.
Claude, Cursor, Codex, and Grok Build can read a SKILL.md you add from the publisher's repository. Grok Bot installs it when you send the skill's GitHub folder link in a chat.