IMSI catcher and downgrade
5G NRA rogue base station or a forced downgrade to 2G exposes subscriber identity and traffic metadata.
Every protocol in the catalog has a threat model: how it is attacked in practice, what the defender can observe, and which control removes the precondition. Scenarios are conceptual and run entirely in the browser.
Defensive education only. These are conceptual models for defenders: no exploit code, payloads or tooling. Never test attacks against systems you do not own or have explicit written permission to assess.
Showing 80 of 80 attack scenarios
A rogue base station or a forced downgrade to 2G exposes subscriber identity and traffic metadata.
Control of the HTTP-01 path or DNS zone lets an attacker obtain a certificate for a domain they do not own.
Broad vhost permissions or the default guest account let a client publish to or consume from queues it should not touch.
Gratuitous ARP replies bind the gateway IP to the attacker MAC, placing the attacker on-path.
BACnet has no native authentication, so anyone who reaches UDP 47808 can write setpoints and schedules.
Forged BFD control packets can keep a dead path looking alive or flap a healthy one.
A network announces a prefix it does not own, and neighbours prefer the more specific or shorter path.
If pairing falls back to Just Works, an on-path attacker can pair with both sides without a shared secret.
Any node on a CAN bus can transmit frames, so a compromised ECU can spoof safety-critical messages.
Spoofed CoAP requests to discovery or large-resource endpoints generate responses aimed at a victim.
An unauthorised DHCP server answers first and hands out a malicious gateway or DNS server.
Short DKIM keys can be factored, and a captured signed message can be replayed if the body is unchanged.
A p=none policy or a missing sp= record leaves subdomains available for spoofing.
A forged answer that matches the transaction ID and source port is cached and served to later clients.
Stripping the DO bit or querying NSEC records can bypass validation or enumerate a zone.
A flood of frames with random source MAC addresses overflows the switch forwarding table, so unknown-unicast traffic is broadcast to every port.
FTP sends credentials and data in the clear, and the PORT command can be abused to reach other hosts.
Server reflection and missing authorisation expose RPC methods that clients were never meant to call directly.
Conflicting Content-Length and Transfer-Encoding headers make a proxy and origin disagree about where one request ends.
Highly compressed header blocks expand into enormous header tables, consuming server memory.
Early data captured from a previous connection is replayed to a server that accepts 0-RTT.
Forged redirects convince a host to send traffic through an attacker-chosen next hop.
A flood of EAPOL start frames overwhelms the authenticator and prevents legitimate supplicants from authenticating.
Forged IGMP queries and joins create multicast forwarding state that floods a segment.
Passwords leaked from another breach are replayed against IMAP, which often lacks MFA.
A PSK hash captured from aggressive mode can be attacked offline, and a weak secret falls quickly.
Packets carry a forged source address, hiding the sender and enabling reflection attacks.
A forged Router Advertisement installs the attacker as the default gateway or DNS server for the segment.
A rogue router forms an adjacency and floods link-state PDUs that distort the topology.
A Kafka broker without authentication exposes topics, consumer offsets and produced records.
Any authenticated user can request a service ticket for an account with an SPN and crack its password offline.
A device that speaks LACP can join a link aggregation group and receive a share of the bundled traffic.
Unescaped input in an LDAP filter can change the query logic and bypass authentication or expose entries.
Forged LLDP TLVs advertise a fake switch or phone identity, poisoning topology and inventory records.
Keys extracted from a device allow an attacker to derive session keys and inject uplinks as that device.
A short commissioning window and a weak setup passcode can allow an attacker to commission a device into their fabric.
A forged multicast answer maps a local service name to the attacker address.
Modbus has no authentication, so anyone who can reach port 502 can write coils or registers.
A misconfigured label stack or route target lets traffic from one VPN reach another.
A client that can subscribe to wildcards or publish to another device topic can forge sensor data or commands.
An open NATS client port lets anyone publish or subscribe to subjects used by the application.
Reused SSH credentials or an over-privileged account allows a NETCONF client to change device configuration.
Two devices forward APDUs in real time so a reader believes the card is present when it is far away.
NFS with AUTH_SYS trusts the client to state its UID and GID, so a modified client can claim another identity.
Forged NTP replies shift a client clock, while legacy monlist queries amplify traffic.
A leaked bearer token is replayed until it expires, and broad scopes widen the damage.
An OPC UA server that allows anonymous sessions exposes tags and methods to anyone who can reach it.
A loose redirect URI allowlist sends the authorization code to an attacker-controlled endpoint.
A rogue router forms an adjacency and floods false link-state advertisements into the area.
POP3 without TLS sends usernames and passwords in plaintext across the local network.
A connection ID that never changes lets an observer link a user across networks and sessions.
A weak RADIUS shared secret allows forged access requests or response tampering.
Redis bound to a reachable interface without a password allows arbitrary commands, including writes and module loading.
A forged RIP update advertises a shorter route for a prefix and attracts traffic to the attacker.
An attacker who guesses the RTP ports and SSRC can inject audio or video into a call.
A forged assertion is inserted beside the signed one so the service provider validates the signature but reads the attacker attributes.
A flood of spoofed INIT chunks consumes bandwidth and association table resources.
A misconfigured SFTP server or stolen credentials allow access outside the intended directory.
Weak digest credentials allow an attacker to register as a user and route calls through the attacker.
Coerced authentication is relayed to another SMB server that accepts NTLM, granting the attacker the victim session.
A misconfigured mail server relays mail for anyone or accepts a forged sender identity.
Devices left on the default public or private community expose configuration and statistics.
A neutral all-qualifier or missing record lets anyone send mail claiming to be the domain.
Automated guessing against exposed SSH services eventually finds a weak password.
Unsanitised CRLF sequences in STOMP headers can create extra headers or split frames.
Faster BPDUs from an attacker make it the root bridge, pulling inter-switch traffic through its link.
Half-open connections exhaust the server backlog so legitimate clients cannot complete the handshake.
Telnet carries credentials and commands in plaintext, so an on-path attacker can read or take over a session.
An on-path attacker strips modern TLS or presents a certificate the client should reject.
Spoofed requests to open UDP services produce large responses aimed at the victim.
A frame with two 802.1Q tags can cross a trunk port when the native VLAN is left untagged.
Forged advertisements with a higher priority make the attacker the virtual router master.
A host that can send VXLAN frames injects traffic into another VNI on the same underlay.
A writable WebDAV share accepts PUT requests and serves the uploaded content back to other users.
ICE candidates can reveal local and public IP addresses to the remote peer before consent.
A malicious page opens a WebSocket to a victim site and relies on the browser sending session cookies.
A rogue access point reuses the corporate SSID and captures clients that associate with the stronger signal.
A leaked private key lets an attacker impersonate a peer and use its allowed IPs.
A controller that accepts non-S2 inclusion lets an attacker join or impersonate a device with weaker security.
Many Zigbee networks still use the well-known default trust-centre key, so an attacker can join and derive network keys.