CRHCSEJul 16

Setup Complete, Now You Are Compromised: Weaponizing Setup Instructions Against AI Coding Agents

arXiv:2607.151439.0
Predicted impact top 38% in CR · last 90 daysOriginality Incremental advance
AI Analysis

For developers and users of AI coding agents, this work reveals a critical security blind spot in how agents handle setup instructions, showing that current harness-model combinations are insufficient to prevent package-install-time attacks.

AI coding agents blindly install dependencies from project documentation without verifying names, sources, or vulnerabilities, making documentation a vector for supply-chain attacks. Across 12 scenarios in 5 attack classes, agents catch blatant typosquats but miss plausible separator-confusion names and source-based attacks like registry redirection; a deterministic pre-install check closes most of the gap.

AI coding agents set up projects by reading documentation and installing the dependencies it lists, without verifying their names, sources, or known vulnerabilities. By editing only a README, requirements file, or Makefile, an attacker can redirect the agent to an untrusted registry, a known-vulnerable version, or a wrong-but-plausible name: documentation becomes a vector for code execution. We present the first systematic evaluation of package-install-time supply-chain attacks delivered through ordinary project-setup documentation across production coding-agent harnesses, probing frontier models on twelve scenarios in five attack classes, grounded in documented incidents. The same model catches an attack through one harness and installs it through another: install-time security rests on the harness-model combination, not the model alone. Agents catch blatant typosquats reliably, but plausible separator-confusion names (azurecore for azure-core) slip through, and how often depends on the harness-model pairing. Source-based attacks like registry redirection are missed almost everywhere. The source blind spot recurs on npm and Cargo, where nearly every model installs the untrusted dependency; name detection carries over less consistently across ecosystems. Security-oriented prompts recover part of the gap but only for the dimension they name; a deterministic pre-install check that verifies names, sources, and versions before any code runs closes most of it.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes