Surviving the Router: Optimizing Skill Injections for Retrieval and Execution
To address this limitation, we introduce CORSA (Cluster Optimization for Router-Aware Skill Attacks), a router-aware attack that optimizes skill injections for both retrieval and execution across clusters of related tasks.
ProofPaper ↗
Key points
- AI agents increasingly rely on modular third-party "skills" that are dynamically selected by skill routers to execute complex tasks.
- While recent studies highlight the threat of prompt injections embedded in these skills, existing evaluations often assume settings where the malicious skill is already selected for execution.
- We show that this assumption can substantially overestimate attack success.
- We evaluate skill injection attacks under router-managed multi-skill settings by extending the benchmark introduced by SkillRouter with eight malicious payload categories.
Sources (1)
- [1]Surviving the Router: Optimizing Skill Injections for Retrieval and ExecutionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 10:28 AM
To address this limitation, we introduce CORSA (Cluster Optimization for Router-Aware Skill Attacks), a router-aware attack that optimizes skill injections for both retrieval and execution across clusters of related tasks.
AI agents increasingly rely on modular third-party "skills" that are dynamically selected by skill routers to execute complex tasks.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 6, 2026Can AI Agents Make Open-Ended Scientific Discovery? Evidence from Station
- Oct 5, 2026MCP for agent-to-agent comms may be the riskiest protocol you've never heard of
- Oct 5, 2026Sharing AI progress in mathematics
- Sep 30, 2026Last Week in AI #345 - 5 new models, 9 misalignment incidents, some Dots
- Sep 29, 2026Claude Code’s Next Era — Thariq Shihipar, Anthropic
- Jun 16, 2026Securing the future of AI agents