Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add dslsdzc/rev-skills --skill re-teegit clone --depth 1 https://github.com/dslsdzc/rev-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/dslsdzc/rev-skills/re-tee)<a href="https://agentmods.dev/skills/dslsdzc/rev-skills/re-tee"><img src="https://agentmods.dev/badge/skills/dslsdzc/rev-skills/re-tee/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/dslsdzc/rev-skills/re-tee"><img src="https://agentmods.dev/badge/skills/dslsdzc/rev-skills/re-tee.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00073 | $0.04557 |
| Opus 5 | $0.00036 | $0.02278 |
| Sonnet 5 | $0.00015 | $0.00911 |
| Haiku 4.5 | $0.00007 | $0.00456 |
Grade A, and why
re-tee scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.
TEE / TrustZone 逆向（OP-TEE / Trusted App / secure storage）
何时使用 / 何时不用
- 用：TrustZone/OP-TEE 类 TEE 分析——可信应用（Trusted App）逆向、命令分发表还原、secure storage（安全存储/设备密钥）对象与数据流
- 用：设备密钥/信任根方向——DRM 硬件信任根（[[re-drm]] 的 L1 类）、设备密钥提取与保护机制（授权研究）
- 用：TEE 接口面——SMC 调用、主机侧 TEE 驱动 ioctl、client 库调用序列还原
- 用：固件安全评估——评估目标设备 TEE 实现的接口暴露面、存储保护强度与信任根边界
- 不用：普通 App 层（走 [[re-mobile]]）；Windows 内核/驱动（走 [[re-kernel]]，TEE 侧是 Linux/ARM 域）
- 不用：纯固件提取解包（先走 [[re-fw-extract]]，本技能在其后分析 TEE 组件）
- 不用：目标只是普通加密数据（走 [[re-crypto-*]] 系列，无 TEE 组件时不必进本技能）
- 不用：仅需判定「设备是否用了 TEE」（无镜像、无调用面可分析）——特征级证据即可，不进深度流程
- 不用：EL2 hypervisor/虚拟化域（走 [[re-hypervisor]]，TEE 是 EL1/EL3 域）
- 注意：secure world 内动态调试通常不可行（见坑 3）——默认静态分析 + 主机侧观察；主机侧动态执行按 [[platform-tips]] 最高原则在沙箱内进行
工具准备
静态分析（镜像解析/反编译）免沙箱；主机侧动态（跑 client、hook ioctl）按 [[platform-tips]] 最高原则进沙箱；secure world 内动态不做（见坑 3），默认以静态 + 主机侧观察为主。
反编译工作台（[[re-ghidra]] / [[re-ida]]，ARM64）
- [[re-ghidra]]（默认）：导入 TEE OS 镜像与 TA（.ta 需先切掉签名头，见步骤 2 与坑 5）
- [[re-ida]]：备选；两者均需 ARM64 支持
- 验证: 能反编译 ARM64 代码，定位
smc指令调用点与TA_InvokeCommandEntryPoint类分发表
固件提取工具链（[[re-fw-extract]]）
- binwalk/unblob、magic 扫描、字节序判断——从 bootrom/启动链/设备固件中定位并提取 TEE OS 与 TA 镜像
- 验证:
binwalk --version（安装命令见 [[re-fw-extract]]「工具准备」）
python3（二进制解析/脚本）
- Linux:
apt install python3/dnf install python3/pacman -S python；macOS: 自带；Windows: 官方安装器或choco install python - 验证:
python3 --version - 用途: 签名头/ELF 载荷切分解析、结构字段标注、解密还原脚本（衔接 [[re-crypto-decrypt]]）
参考实现（OP-TEE 官方开源，理解结构用）
- 官方仓库 optee_os / optee_client / optee_examples（与 qemu 仿真环境）提供 TEE 结构与调用序列的权威参照——分析目标与 OP-TEE 差异时按官方源码核对，不凭经验值
- qemu 仿真环境可跑通 client → TA 全链路调用，用于对照调用序列与参数布局；仿真与真机的差异（真实驱动/中断/存储介质）以真机为准
frida / 主机侧观察（[[re-frida]]）
- 主机侧动态：hook client 库调用序列与 ioctl 参数（frida 安装与沙箱原则见 [[re-frida]]，动态执行默认沙箱）
- 验证:
frida --version
操作步骤
按顺序执行，每步产物（镜像定位、头结构标注、命令号表）记录证据路径 + sha256（见 [[re-triage]]），供报告引用。
- 架构定位（world 划分 / SMC / TEE OS 入口）：
- TrustZone 划分 normal world（普通世界，NS）与 secure world（安全世界，S）；ARMv8-A 中 TEE OS 通常运行在 secure EL1，可信应用在 secure EL0，EL3 监控层（monitor）承载安全监控固件（如 ATF BL31 类启动固件，泛化）
- SMC 指令是 normal world 进入 secure world 的通道：AArch64 为
smc #imm（AArch32 为smc），从 EL1/EL2 执行即陷入 EL3 监控层，监控层再分发给 TEE OS；反编译中搜smc指令即调用面落点 - SMC 参数约定（SMCCC）：功能号在 w0（含服务/调用约定标识位），参数在 w1-w7（SMC32）或 x1-x17（SMC64），返回值从 w0 起——标注调用点按约定表读寄存器，不按普通 ABI 猜（见坑 4）
- 特征侦查：版本字符串、导出 API 字符串表、SMC 功能号常量表先 grep 定位——字符串表往往直接给出 TEE OS 类型与版本，比逐个反编译快
- 定位 TEE OS 加载入口：bootrom → 启动固件 → TEE OS 镜像（镜像头 magic、加载/验签代码）——沿启动链搜 TEE 镜像特征（镜像头魔数、版本字符串、导出 API 字符串表）；TEE OS 也可内嵌于设备固件分区（[[re-fw-extract]] 提取后定位）
- 产物：world/EL 划分图 + TEE OS 镜像位置与加载入口
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 121 lines · 73 tokens per session scan A 4795ad063965
re-tee is a skill published in the GitHub repository dslsdzc/rev-skills (54 stars, last pushed 14d ago), licensed Apache-2.0. It adds 73 tokens to every session and 4,557 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
analyzing-golang-malware-with-ghidra
Reverse engineer Go-compiled malware in Ghidra by parsing Go buildinfo and pclntab structures, recovering stripped/obfuscated function names (e.g. via GoResolver), and extracting embedded module/dependency strings and types from Go binaries. Use when analyzing a Go-language malware sample, deobfuscating a…
reverse-engineering-arm-binaries
Reverse engineers ARM/AArch64 malware by identifying the architecture and instruction set state (ARM/Thumb), parsing ELF/Mach-O ARM headers, and orienting analysis around the ARM calling convention. Activates for requests to reverse ARM binaries, analyze AArch64 malware, or handle ARM/Thumb instruction-set decoding.
analyzing-golang-malware-with-ghidra
Reverse engineer Go-compiled malware in Ghidra by parsing Go buildinfo and pclntab structures, recovering stripped/obfuscated function names (e.g. via GoResolver), and extracting embedded module/dependency strings and types from Go binaries. Use when analyzing a Go-language malware sample, deobfuscating a…
analyzing-golang-malware-with-ghidra
Reverse engineer Go-compiled malware in Ghidra by parsing Go buildinfo and pclntab structures, recovering stripped/obfuscated function names (e.g. via GoResolver), and extracting embedded module/dependency strings and types from Go binaries. Use when analyzing a Go-language malware sample, deobfuscating a…
Reverse Engineering & Binary Analysis
Binary analysis, assembly interpretation, disassembly, decompilation, firmware RE, and protocol reverse engineering.
ot-ics-attack
A guide to assessing operational technology and industrial control systems such as SCADA, DCS, and PLC environments. These systems monitor and control physical processes and often use specialized network protocols.