ALARA for Agents: Least-Privilege Context Engineering Through Portable Composable Multi-Agent Teams
Christopher J. Agostino 1 , Nayan D’Souza 1 2 · 2026
Read ALARA for Agents: Least-Privilege Context Engineering Through Portable Composable Multi-Agent Teams by Christopher J. Agostino 1 , Nayan D’Souza 1 2 (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-2603-20380.txt.
Description
Agostino and D'Souza apply a radiation safety idea to managing multi-agent AI teams. The ALARA rule keeps exposure as low as reasonably achievable. Their version keeps agent context lean through a plain-text data layer. Users declare each agent's tool access directly and can edit the tools. A shell program, npcsh, loads the team and runs the agents. Tests covered 22 local models on 115 practical tasks. The tasks spanned file work, web search, scripting, and delegation.
Catalog record
| CREATOR | Christopher J. Agostino 1 , Nayan D’Souza 1 2 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28033 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.20380