# How Your Credentials Are Leaked by LLM Agent Skills: An Empirical Study

arXiv: https://arxiv.org/abs/2604.03070 · HTML: https://mlsanity.com/paper/2604.03070/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 16 of its 57 references that are in the map.
- First posted on arXiv 3 Apr 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, 16 of 16)

- [Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection](https://mlsanity.com/paper/2302.12173/) (2023)
- [Prompt Injection Attack to Tool Selection in LLM Agents](https://mlsanity.com/paper/2504.19793/) (2025)
- [IsolateGPT: An Execution Isolation Architecture for LLM-Based Agentic Systems](https://mlsanity.com/paper/2403.04960/) (2024)
- [Agent Skills in the Wild: An Empirical Study of Security Vulnerabilities at Scale](https://mlsanity.com/paper/2601.10338/) (2026)
- [Imprompter: Tricking LLM Agents into Improper Tool Use](https://mlsanity.com/paper/2410.14923/) (2024)
- [MCP Safety Audit: LLMs with the Model Context Protocol Allow Major Security Exploits](https://mlsanity.com/paper/2504.03767/) (2025)
- [AttriGuard: Defeating Indirect Prompt Injection in LLM Agents via Causal Attribution of Tool Invocations](https://mlsanity.com/paper/2603.10749/) (2026)
- [Small World with High Risks: A Study of Security Threats in the npm Ecosystem](https://mlsanity.com/paper/1902.09217/) (2019)
- [Agent Skills Enable a New Class of Realistic and Trivially Simple Prompt Injections](https://mlsanity.com/paper/2510.26328/) (2025)
- [Agents of Chaos](https://mlsanity.com/paper/2602.20021/) (2026)
- [Prompt Injection Attacks on Agentic Coding Assistants: A Systematic Analysis of Vulnerabilities in Skills, Tools, and Protocol Ecosystems](https://mlsanity.com/paper/2601.17548/) (2026)
- [Formal Analysis and Supply Chain Security for Agentic AI Skills](https://mlsanity.com/paper/2603.00195/) (2026)
- [MalTool: Malicious Tool Attacks on LLM Agents](https://mlsanity.com/paper/2602.12194/) (2026)
- [What are Weak Links in the npm Supply Chain?](https://mlsanity.com/paper/2112.10165/) (2021)
- [SoK: Analysis of Software Supply Chain Security by Establishing Secure Design Properties](https://mlsanity.com/paper/2406.10109/) (2024)
- [GoSurf: Identifying Software Supply Chain Attack Vectors in Go](https://mlsanity.com/paper/2407.04442/) (2024)

## Built on by (most cited first, 19 of 19)

- [When Safe Skills Collide: Measuring Compositional Risk in Agent Skill Ecosystems](https://mlsanity.com/paper/2606.00448/) (2026)
- [AgentTrap: Measuring Runtime Trust Failures in Third-Party Agent Skills](https://mlsanity.com/paper/2605.13940/) (2026)
- [Cloak and Detonate: Scanner Evasion and Dynamic Detection of Agent Skill Malware](https://mlsanity.com/paper/2607.02357/) (2026)
- [RouteGuard: Internal-Signal Detection of Skill Poisoning in LLM Agents](https://mlsanity.com/paper/2604.22888/) (2026)
- [Safety Testing LLM Agents at Scale: From Risk Discovery to Evidence-Grounded Verification](https://mlsanity.com/paper/2607.01793/) (2026)
- [Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries](https://mlsanity.com/paper/2607.10113/) (2026)
- [Trust Me, Import This: Dependency Steering Attacks via Malicious Agent Skills](https://mlsanity.com/paper/2605.09594/) (2026)
- [SUDP: Secret-Use Delegation Protocol for Agentic Systems](https://mlsanity.com/paper/2604.24920/) (2026)
- [SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents](https://mlsanity.com/paper/2606.02302/) (2026)
- [MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill Detection](https://mlsanity.com/paper/2608.19901/) (2026)
- [Covert Assistance: Helpful LLM Agents Evade Oversight in Multi-Agent Systems](https://mlsanity.com/paper/2609.39050/) (2026)
- [SkillPoison: Progressive Skill Poisoning via Successful Experiences](https://mlsanity.com/paper/2610.07645/) (2026)
- [StealthBench: Measuring Operational Stealth in Autonomous Offensive-Security Agents](https://mlsanity.com/paper/2607.26314/) (2026)
- [OBLIVION: Workflow-Level Operational Skill Unlearning for Deployed Agents](https://mlsanity.com/paper/2608.08264/) (2026)
- ["Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables](https://mlsanity.com/paper/2609.35408/) (2026)
- [SkillScreen: High-Recall Screening of Malicious Agent Skills with Decision Models](https://mlsanity.com/paper/2610.17727/) (2026)
- [Encrypted Skills Can Become New Security Exploits If AI Agents Know How to Decipher](https://mlsanity.com/paper/2610.18262/) (2026)
- [Beyond Maliciousness: Static and Runtime Attack-Surface Profiling of Agent Skills](https://mlsanity.com/paper/2610.19264/) (2026)
- [CoRedSkill: Stealthy Skill Composition Attacks via AST-Guided Injection](https://mlsanity.com/paper/2610.20049/) (2026)

## Abstract (arXiv)

Large Language Model (LLM) agents increasingly rely on third-party skills that operate within privileged execution environments and routinely handle sensitive credentials, yet how these credentials are leaked remains largely unexplored. To fill this gap, we present the first large-scale empirical study on credential leakage in agent skills. From 170,226 artifacts on SkillsMP, the largest open-source skill marketplace, we sampled 17,022 skills via stratified random sampling and analyzed each through static secret extraction (regex and AST parsing), dynamic sandbox testing with mock credentials, and cross-referencing developer intent against runtime behavior. Our analysis identifies 520 affected skills containing 1,708 security issues, and yields a taxonomy of 10 leakage patterns. Three findings stand out. First, 76.3% of cases require jointly analyzing natural-language descriptions and programming logic, showing that credential exposure in skills is fundamentally cross-modal. Second, debug logging accounts for 73.5% of vulnerabilities because agent frameworks feed stdout into the LLM context window, turning routine debugging into a credential exposure vector. Third, 89.6% of leaked credentials are immediately exploitable -- 92.5% during routine execution without elevated privileges -- and the fork-based distribution model defeats remediation, as secrets removed from 107 upstream repositories persist across 50+ independent forks. Following responsible disclosure, all malicious skills have been removed and 91.6% of hardcoded cases remediated. We release our dataset, taxonomy, and detection pipeline to support future agent security research.