Supply-Chain Poisoning Attacks Against LLM Coding Agent Skill Ecosystems

TL;DR

This paper uncovers a novel supply-chain attack method, DDIPE, that embeds malicious logic in skill documentation for LLM coding agents, enabling payload execution during normal operations.

Contribution

It introduces DDIPE, a new attack technique exploiting skill documentation to bypass defenses and demonstrates its effectiveness across multiple frameworks and models.

Findings

DDIPE achieves 11.6% to 33.5% bypass rates.

Static analysis detects most attacks, but 2.5% evade detection.

Responsible disclosure led to vulnerabilities being fixed.

Abstract

LLM-based coding agents extend their capabilities via third-party agent skills distributed through open marketplaces without mandatory security review. Unlike traditional packages, these skills are executed as operational directives with system-level privileges, so a single malicious skill can compromise the host. Prior work has not examined whether supply-chain attacks can directly hijack an agent's action space, such as file writes, shell commands, and network requests, despite existing safeguards. We introduce Document-Driven Implicit Payload Execution (DDIPE), which embeds malicious logic in code examples and configuration templates within skill documentation. Because agents reuse these examples during normal tasks, the payload executes without explicit prompts. Using an LLM-driven pipeline, we generate 1,070 adversarial skills from 81 seeds across 15 MITRE ATTACK categories. Across four frameworks and five models, DDIPE achieves 11.6% to 33.5% bypass rates, while explicit instruction attacks achieve 0% under strong defenses. Static analysis detects most cases, but 2.5% evade both detection and alignment. Responsible disclosure led to four confirmed vulnerabilities and two fixes.