Skip to content
KitploitKITPLOIT
StrumentiExploitsBlog
Log in
Invia
StrumentiExploitsBlog
Invia

Strumenti di Hacking, PenTest e Cybersecurity per il tuo Arsenale di Sicurezza!

Kitploit è una directory di strumenti di hacking, cybersecurity e pentesting. Scopri gli ultimi aggiornamenti dei progetti per trovare vulnerabilità, analizzare sistemi, automatizzare i test e rafforzare la tua sicurezza.

FeedContattoPrivacy© 2026 Kitploit

Directory degli strumenti

Categorie

Vedi tutte le categorie
Loading categories
bombini — eBPF Security Monitoring and Sandboxing Agent Based on Aya | Kitploit
Strumenti/GitHubGitHub/bombinisecurity/bombini
Defensive ToolsContainer SecurityIntrusion Detection
GitHubbombinisecurity/bombini

bombini

eBPF Security Monitoring and Sandboxing Agent Based on Aya

Vedi Repository
80111082 giorni faRevisionato da Kitploit
Condividi
Contenuto non disponibile nella lingua richiesta. Visualizzazione della versione inglese.

Bombini: eBPF-based Security Monitoring and Sandboxing Agent

Bombini logo

License CI Book

Bombini is an eBPF-based security agent written entirely in Rust using the Aya library and built on LSM (Linux Security Module) BPF hooks. At its core, Bombini employs modular components called Detectors, each responsible for monitoring and reporting specific types of system events. Detectors with sandbox mode enabled also block forbidden actions right in the kernel.

Why Bombini

  • Filters in the kernel. Rules are compiled into eBPF maps and evaluated by the eBPF programs, so only matching events reach userspace.
  • Blocks in the kernel, not after the fact. Sandbox rules are enforced by LSM BPF hooks: a forbidden exec, file open or connection fails with EPERM instead of being reported once it has already happened.
  • One rule language for detection and enforcement. The same scope/event predicates either report an event or, with sandbox enabled, deny it.
  • Sees what syscall tracing misses. IOUringMon inspects io_uring submissions, and KernelMon reports BPF program and map activity.
  • Easy to ship. A single static musl binary plus BPF objects, available as a container image, a Kubernetes DaemonSet with pod metadata enrichment, or a tarball with a systemd unit.

Quickstart

Bombini needs Linux 6.2+ with BPF LSM enabled. Check the compatibility notes first. The container image is built for x86_64.

Start Bombini with the quickstart config from the repository root:

docker run --pid=host --rm -it --privileged \
  -v $PWD/examples/quickstart:/usr/local/lib/bombini/config:ro \
  -v /sys/fs/bpf:/sys/fs/bpf \
  ghcr.io/bombinisecurity/bombini:v1.2.0

The config consists of three detector files.

procmon.yaml reports setuid to root and forbids executing anything from /tmp:

# Detect privilege escalation to root
setuid:
  enabled: true
  rules:
    - rule: SetUidToRoot
      event: euid == 0

# Block binaries executed from /tmp right in the kernel
bprm_check:
  enabled: true
  sandbox:
    enabled: true
    deny_list: true
  rules:
    - rule: DenyExecFromTmp
      event: path_prefix == "/tmp"

filemon.yaml reports reads of /etc/shadow:

# Detect reads of password hashes
file_open:
  enabled: true
  rules:
    - rule: ShadowRead
      event: path == "/etc/shadow"

netmon.yaml reports connections to the cloud metadata service:

# Detect access to the cloud metadata service: a common way to steal credentials
egress:
  enabled: true
  rules:
    - rule: CloudMetadataAccess
      event: ipv4_dst == "169.254.169.254" AND port_dst == 80

Generate events in another terminal:

$ ./examples/quickstart/trigger.sh
[1] sudo reads /etc/shadow: Setuid and FileOpen events
[2] exec a binary from /tmp: blocked by the sandbox
./examples/quickstart/trigger.sh: 10: /tmp/bombini-quickstart: Operation not permitted
    blocked: exit code 126
[3] connect to the cloud metadata service: Egress event

Bombini prints events to stdout as JSON lines. This is the blocked exec:

{
  "type": "ProcessEvent",
  "process": {
    "start_time": "2026-10-03T14:36:10.256Z",
    "cloned": true,
    "pid": 39988,
    "tid": 39988,
    "ppid": 39984,
    "uid": 503,
    "euid": 503,
    "gid": 1000,
    "egid": 1000,
    "auid": 0,
    "cap_inheritable": "",
    "cap_permitted": "",
    "cap_effective": "",
    "secureexec": "",
    "filename": "dash",
    "binary_path": "/usr/bin/dash",
    "args": "./examples/quickstart/trigger.sh",
    "exec_id": "Mzk5ODg6MjEyNzU3NDAzMjI2MzI2",
    "parent_exec_id": "Mzk5ODQ6MjEyNzU3MzgzNTY5NTc3"
  },
  "parent": { "...": "same fields for the parent process" },
  "blocked": true,
  "process_event": {
    "type": "BprmCheck",
    "binary": "/tmp/bombini-quickstart"
  },
  "timestamp": "2026-10-03T14:36:10.256Z",
  "rule": "DenyExecFromTmp"
}

More ready-made policies, such as GTFOBins shell escapes and privilege escalation, are in examples. The rule library has one file per detector, based on the Falco default rules and Tetragon tracing policies and adapted to Bombini, plus rules only Bombini can express. It runs in audit mode by default: flip sandbox.enabled on a hook to enforce it.

Detectors

DetectorWhat it watchesSandbox
ProcMonProcess execs and exits, setuid/setgid, capabilities, prctl, user namespaces, ptrace✓
FileMonFile open, mmap, truncate, unlink, symlink, chmod, chown, mount, ioctl✓
NetMonIngress and egress TCP connections, socket creation and connect✓
KernelMonBPF map creation, BPF program loads, and access to BPF maps and programs
IOUringMonio_uring submission queue entries
SysEnumMon (experimental)System enumeration: several watched files or binaries touched within a time window

The rule syntax is described in Rules, and the event format in Events.

Installation

  • Container
  • Kubernetes
  • Tarball with systemd
  • Build from source

Full documentation is in the book.

Contributing

Please, check out CONTRIBUTING.md for the contributing guideline.

Scarica lo strumento

Più Popolari

Vedi tutti →

Scopri gli strumenti più utilizzati dalla nostra community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 Il toolkit OSINT definitivo: grafo di indagine, punteggio di affidabilità, famiglie di alias, mappa, timeline e report. 49 moduli, oltre 400 siti, nessuna chiave API richiesta.

    732
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    # App di esecuzione con un tocco GhostLock (CVE-2026-43499)

    506
  • 3

    ARTEX

    baarai/artex

    Piattaforma di penetration testing autonoma multi-agente basata su LLM con agenti planner/worker, grafi di asset e di esplorazione, registrazione del traffico MITM e retesting delle vulnerabilità.

    405
  • 4

    redforge

    cybersecurity91/redforge

    Toolkit red team lato browser con 102 strumenti offline per ricognizione, generazione di payload, attacchi alle credenziali, crittografia, codifica e reportistica dell'engagement.

    342
  • 5

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    Il sistema d'azione per la cybersecurity nativa AI—dove l'intento diventa esecuzione governata, le evidenze diventano memoria operativa, e ogni operazione migliora la successiva.

    261
  • 6

    antifox

    antifox-ai/antifox

    Piattaforma multi-agente di penetration testing autonoma basata sull'IA che esegue ricognizione, scoperta della superficie d'attacco, validazione dello sfruttamento e consegna di report supportati da evidenze.

    259

Esplora tutti gli strumenti

Sfoglia la nostra collezione di strumenti

Vedi tutti gli strumenti →