Protecting Context and Prompts: Deterministic Security for Non-Deterministic AI
Mohan Rajagopalan, Vinay Rao · 2026
Read Protecting Context and Prompts: Deterministic Security for Non-Deterministic AI by Mohan Rajagopalan, Vinay Rao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-10481.txt.
Description
Mohan Rajagopalan and Vinay Rao tackle prompt injection and context tampering in a 2026 paper. Language-model uses inherit inputs from many sources, and older security models cannot verify where those inputs came from or whether they were altered. The paper presents two primitives to close that gap.
Authenticated prompts carry self-contained lineage that can be verified on arrival. Authenticated context chains dynamic inputs with tamper-evident hashes, so any alteration breaks the chain. On top of these, a policy algebra with four proven theorems gives protocol-level resistance even against adversarial agents. Five layered defenses, from resource controls to meaning checks, round out a security design meant to be preventative rather than reactive.
Catalog record
| CREATOR | Mohan Rajagopalan, Vinay Rao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-10481