A structured starter library of threats mapped to Innora research, glossary definitions, product proof pages, and buyer-facing mitigation routes.
3 seed threats
AI Systems
User, tool, document, or retrieved-content instructions alter an AI system decision path and bypass intended guardrails.
AI Systems
Autonomous agents misuse browser, file, API, or workflow tools after prompt manipulation, weak policy checks, or missing approval gates.
AI Systems
Compromised models, datasets, adapters, or inference dependencies introduce backdoors, unsafe behavior, or remote execution paths.
2 seed threats
Software Vulnerabilities
C or C++ parsing, bounds, integer, or lifetime bugs create crashes, denial of service, information disclosure, or code execution opportunities.
Software Vulnerabilities
Malformed vehicle-to-everything protocol input reaches parsing or cryptographic verification before authentication and terminates a network-facing process.
2 seed threats
Web3
Bridge validators, message verification, replay handling, or accounting assumptions fail and let attackers move or mint assets across chains.
Web3
ERC-4337 or similar account-abstraction flows let attackers shift gas, validation, or sponsorship assumptions into exploitable paymaster behavior.
2 seed threats
Threat Intelligence
Nation-state or organized intrusion sets gain footholds through exposed services, supply chain access, credential theft, phishing, or trusted tooling.
Threat Intelligence
Ransomware operators retain access, disable controls, stage payloads, and create recovery blockers before encryption or extortion.
2 seed threats
Mobile and Reverse Engineering
Runtime instrumentation or weak certificate validation lets authorized testers or attackers inspect traffic that the app expected to protect.
Mobile and Reverse Engineering
Static and dynamic analysis recover string, control-flow, or anti-debugging behavior from protected applications during authorized review.