Vai al contenuto
New

DeepSeek V4.1 Flash free for 7 days: 10M tokens a day after a top-up of $10 or more. Until Nov 30.

How it works
aeondave/emulation-techniqueMobile, firmware e hardware

emulation-technique

⁨Auth/lab: generic emulation methodology for binaries, firmware, kernels, and MCU/board targets; static-to-dynamic routing, tool choice, validation signals.⁩

v2026.11.13-21cd5683 installazioni+3 in 7 giorni39 StelleMITsincronizzato 10 ore fa

Installa

Non hai ancora la CLI di NotHoly? Installala prima.

$ curl -fsSL https://notholy.com/install.sh | sh
$ notholy skills install aeondave/emulation-technique

Scrive in ⁦~/.notholy/skills⁩ dopo aver controllato la firma e lo SHA-256.

FirmaFirmato, ed25519
p+wpxIlEA0…2OVjOCCw==La CLI controlla lo SHA-256 e la firma prima di scrivere qualsiasi file.
Pacchettoe0d6d86d…355e83fa
e0d6d86de504…feca355e83fa5 file · ⁦8.4 KB⁩
ScansioneBassonessuna segnalazione
La scansione non ha trovato nulla da segnalare in questa versione.Vedi la scansione completa
Sorgentegithub.com/AeonDave/malskill/tree/main/offensive-techniques/emulation-technique@21cd568

Questa voce non è ancora scritta in italiano, quindi è mostrata in un'altra lingua: English.

Use emulation to answer a runtime question with the smallest faithful environment that can produce evidence.

When evaluating a suspected vulnerability, pair with vuln-research for the research decision and its binary-experiment reference for distinguishing target behavior from harness artifacts.

When this skill applies

  • You have a binary, firmware image, rootfs, kernel, bootloader, shellcode, driver, MCU image, or unknown architecture artifact.
  • Static analysis found architecture, format, APIs, strings, init scripts, vector tables, MMIO, network services, unpacking, decryption, or anti-analysis behavior, but runtime evidence is needed.
  • You need to decide between native execution, user-mode emulation, OS-layer emulation, full-system emulation, board simulation, CPU-only emulation, or hardware.

Safety and scope gates

  • Treat every dynamic run as untrusted execution. Use an isolated VM/container, snapshots, no shared secrets, controlled networking, and localhost-only forwards unless scope explicitly permits more.
  • Do not run unknown malware or firmware on the analyst host natively.
  • Preserve the original artifact. Work from copies and record hashes of every input and modified output.
  • Prefer read-only mounts and explicit scratch directories until the workflow requires mutation.
  • Do not expose guest services broadly. Bind forwards to localhost and capture traffic/transcripts as evidence.

Static-to-dynamic gate

Do not emulate blindly. Before the first run, record:

  • Artifact type: ELF, PE, Mach-O, raw blob, shellcode, kernel, rootfs, disk image, bootloader, MCU firmware, UEFI, archive, filesystem.
  • CPU and ABI: arch, bits, endian, ARM/Thumb mode, hard/soft float, libc, interpreter, OS ABI, privilege level.
  • Loader facts: entry point, base/load address, program headers, imports, required libraries, rootfs or DLL/framework dependency set.
  • Environment facts: config paths, expected devices, NVRAM, /proc, registry, network ports, UARTs, GPIO/JTAG/SWD, MMIO ranges, DTB/SVD clues.
  • Runtime question: what output, branch, API call, packet, decrypted data, crash, console line, service response, or memory/register state will prove progress.

Escalate to dynamic analysis when static analysis cannot answer behavior, unpacking, anti-analysis, runtime config, syscall/API effects, service binding, interrupt/peripheral behavior, or kernel/driver reachability.

Choose the lane

Target need First dynamic lane Public tools
Isolated decode loop, shellcode, custom VM handler, no OS CPU-only emulation Unicorn, Ghidra emulator, radare2 ESIL
Linux/BSD user-mode executable with known sysroot Process emulation QEMU user-mode, strace, ltrace, GDB
PE/Mach-O/ELF with OS APIs to hook or fake OS-layer emulation Qiling, Frida, debugger API hooks
Router/camera/IoT Linux firmware services Firmware rehosting binwalk, unsquashfs, ubi_reader, QEMU user/system, FirmAE/Firmadyne
Kernel, initramfs, disk, driver, bootloader, UEFI Full-system emulation QEMU system, OVMF/SeaBIOS, GDB remote, qemu-img
MCU/RTOS/board firmware with MMIO, UART, GPIO, timers Board simulation Renode, QEMU system board models, OpenOCD/GDB, SVD-aware RE
Peripheral parity is missing but routine is clear Partial model Unicorn/Qiling hooks, Renode custom platform, MMIO stubs
Emulator cannot model the target and the objective requires hardware behavior Physical lab UART/JTAG/SWD/SPI tools, logic analyzer

Pick the weakest lane that can answer the question. If a lane fails because it lacks OS, syscalls, files, devices, interrupts, or peripherals, move one level closer to the real environment.

Core loop

  1. Triage: identify format, arch, dependencies, and expected environment.
  2. Hypothesis: state exactly what runtime behavior you expect and how you will observe it.
  3. Minimal run: use the narrowest emulator setup with timeouts and bounded output.
  4. Instrument: add tracing only where needed: syscalls, APIs, blocks, memory, UART, GDB, network, MMIO.
  5. Patch or stub: fake missing environment only when the missing piece is not the behavior under study.
  6. Validate: capture one replayable signal: console line, response, branch hit, memory dump, register state, trace, packet, crash, or file mutation.
  7. Escalate or stop: if the signal answers the question, stop. If not, update the missing-environment model and rerun.

Evidence standards

  • "It started" is not enough. Record what executed, what was observed, and why it proves the claim.
  • For services, prove a bound endpoint and one protocol-level interaction with external tooling.
  • For firmware, preserve extraction logs, init path, architecture, chosen emulator command, and network/UART transcript.
  • For MCU/board work, distinguish CPU instruction execution from board/peripheral parity.
  • For kernel work, capture boot banner, panic/oops text, GDB stop, or driver interaction evidence.
  • For hooks and patches, label every fake response and do not confuse it with native target behavior.

Common failure pivots

  • exec format error or illegal instruction: wrong arch, endian, ABI, CPU extension, ARM/Thumb mode, or loader.
  • Missing interpreter/library: build or extract a sysroot; set -L, QEMU_LD_PREFIX, LD_LIBRARY_PATH, or Qiling rootfs.
  • Missing syscall/API: hook or stub in Qiling/Frida, or switch to QEMU/full-system.
  • Hangs in MMIO loop: identify address range; use Renode/custom peripheral model or stub reads/writes.
  • No serial output: wrong console device, baud irrelevant in virtual UART, wrong DTB, wrong machine, guest waiting on storage or init.
  • Service unreachable: wrong guest IP, daemon did not start, firewall/init script missing, port forward bound to wrong NIC, service needs NVRAM/config.
  • Dynamic run changes behavior: anti-debug/anti-VM/time/network checks; patch or emulate the check explicitly and record it.

Resources

  • references/static-to-dynamic-gates.md — load before choosing dynamic execution when the target class is still uncertain.
  • references/userland-emulation.md — Linux/BSD user-mode, OS-layer emulation, sysroots, syscall/API tracing, and service processes.
  • references/firmware-rehosting.md — Linux firmware/rootfs extraction, init/service rehosting, NVRAM/config grafting, and network validation.
  • references/board-and-kernel-emulation.md — MCU/RTOS, Renode/QEMU board choice, Linux kernel boot/debug, DTB/SVD/MMIO decisions.

File di questa versione

Permessi

PermessoDichiaratoTrovato dalla scansione
Legge fileNoNo
Scrive fileNoNo
Esegue comandi shellSìSì
Usa la reteSìSì
Usa i connettoriNoNo
Controlla un browserNoNo

Segnalazioni della scansione

Rivisto il 13 nov 2026

La scansione non ha trovato nulla da segnalare in questa versione.

Una scansione pulita abbassa il rischio, ma non prova che il pacchetto sia sicuro. Leggi quello che installi.

VersionePubblicataCommitSHA-256FirmaCambio di permessi
v2026.11.13-21cd56813 nov 202621cd568e0d6d86dFirmata–

Chiedi tutto quello che il tuo perimetro consente.

Crea un account, aggiungi crediti e fai la prima domanda in chat o nel terminale.

  1. 1/3

    Crea l'account

    Privato o azienda. Puoi passare ad azienda in seguito senza perdere saldo o chiavi.

    PrivatoAzienda
  2. 2/3

    Aggiungi crediti

    Con carta. Paghi a token, senza abbonamento.

    $5.00ricarica minima
  3. 3/3

    Chiedi

    In chat, oppure installa la CLI ed esegui notholy login nel terminale.

    $ notholy login