# A Finger on the Scale: Covert Policy Steering through Agentic Skills

arXiv: https://arxiv.org/abs/2609.02564 · HTML: https://mlsanity.com/paper/2609.02564/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 10 of its 70 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 10 of 10)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [StruQ: Defending Against Prompt Injection with Structured Queries](https://mlsanity.com/paper/2402.06363/) (2024)
- [R-Judge: Benchmarking Safety Risk Awareness for LLM Agents](https://mlsanity.com/paper/2401.10019/) (2024)
- [Prompt Injection Attack to Tool Selection in LLM Agents](https://mlsanity.com/paper/2504.19793/) (2025)
- [MPMA: Preference Manipulation Attack Against Model Context Protocol](https://mlsanity.com/paper/2505.11154/) (2025)
- [SkillAttack: Automated Red Teaming of Agent Skills through Attack Path Refinement](https://mlsanity.com/paper/2604.04989/) (2026)
- [Bias Beware: The Impact of Cognitive Biases on LLM-Driven Product Recommendations](https://mlsanity.com/paper/2502.01349/) (2025)
- [Unveiling and Mitigating Bias in Large Language Model Recommendations: A Path to Fairness](https://mlsanity.com/paper/2409.10825/) (2024)
- [Trust Me, Import This: Dependency Steering Attacks via Malicious Agent Skills](https://mlsanity.com/paper/2605.09594/) (2026)

## Abstract (arXiv)

Reusable agent skills extend large language model (LLM) agents with task procedures, tool-use guidance, and output constraints. Yet these skills also act as externalized behavioral policies, which create a supply-chain risk: a third-party skill may preserve the declared task and valid output interface while covertly redirecting agent decisions toward an undisclosed objective. We formalize Skill Policy Integrity, which requires a Skill-induced policy to remain aligned with its declared functionality and the user-authorized objective. We further present SkillShift, a constrained black-box framework for covert policy steering without explicit target command injection or task hijacking. It combines semantically plausible policy edits with hierarchical validation, failure-guided optimization, and strategy compression to preserve effectiveness, output validity, transferability, and inconspicuousness. We instantiate this threat in agentic commerce and software dependency use, with SkillShift achieving attacker-favored selection rates of 81.33% and 63.33% while maintaining a 100% utility-preserving rate. The frozen policies also transfer without further optimization across heterogeneous LLM backends and agent environments. Moreover, the evaluated scanners fail to detect the constructed skills, motivating behavioral auditing of reusable skills as agent policy artifacts.