Skill-Inject Paper on Agent Skill Prompt Injection Accepted at NeurIPS 2026

The Skill-Inject benchmark, which measures how resistant coding agents like Claude Code are to malicious instructions hidden in agent skill files, has been accepted at NeurIPS 2026, with the work already cited over 80 times.

2026-09-25 ~ 2026-09-25 · 2 related posts

• Skill-Inject Benchmark Shows Frontier Agents Fall for Malicious Skills, Accepted at NeurIPS 2026 — maksym_andr · 2026-09-25
• Skill-Inject on prompt injection in agent skill files accepted at NeurIPS 2026, 80+ citations — maksym_andr · 2026-09-25