# Labels Are Not Endpoints: Treatment Leakage and Construct Validity in MCP Agent Security Evaluation

arXiv: https://arxiv.org/abs/2608.12880 · HTML: https://mlsanity.com/paper/2608.12880/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 32 references that are in the map.
- First posted on arXiv 13 Aug 2026.

## Builds on (most cited first, 13 of 13)

- [Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection](https://mlsanity.com/paper/2302.12173/) (2023)
- [InjecAgent: Benchmarking Indirect Prompt Injections in Tool-Integrated Large Language Model Agents](https://mlsanity.com/paper/2403.02691/) (2024)
- [Agent Security Bench (ASB): Formalizing and Benchmarking Attacks and Defenses in LLM-based Agents](https://mlsanity.com/paper/2410.02644/) (2024)
- [Measuring what Matters: Construct Validity in Large Language Model Benchmarks](https://mlsanity.com/paper/2511.04703/) (2025)
- [Metamorphic Testing: A New Approach for Generating Next Test Cases](https://mlsanity.com/paper/2002.12543/) (2020)
- [MCP Security Bench (MSB): Benchmarking Attacks Against Model Context Protocol in LLM Agents](https://mlsanity.com/paper/2510.15994/) (2025)
- [MCP-SafetyBench: A Benchmark for Safety Evaluation of Large Language Models with Real-World MCP Servers](https://mlsanity.com/paper/2512.15163/) (2025)
- [MCP-Guard: A Multi-Stage Defense-in-Depth Framework for Securing Model Context Protocol in Agentic AI](https://mlsanity.com/paper/2508.10991/) (2025)
- [VIGIL: Defending LLM Agents Against Tool Stream Injection via Verify-Before-Commit](https://mlsanity.com/paper/2601.05755/) (2026)
- [MCP-ITP: An Automated Framework for Implicit Tool Poisoning in MCP](https://mlsanity.com/paper/2601.07395/) (2026)
- [How Adversarial Environments Mislead Agentic AI?](https://mlsanity.com/paper/2604.18874/) (2026)
- [SafeMCP: Proactive Power Regulation for LLM Agent Defense via Environment-Grounded Look-Ahead Reasoning](https://mlsanity.com/paper/2606.01991/) (2026)
- [ShareLock: A Stealthy Multi-Tool Threshold Poisoning Attack Against MCP](https://mlsanity.com/paper/2606.27027/) (2026)

## Abstract (arXiv)

Security evaluations of tool-using agents often equate stored labels with behavioral facts. We audit a preserved campaign by tracing 10,200 execution rows to 180 model-bound requests, 45 semantic requests, and 15 observable stimuli. Two schema treatments were delivered, but the planned external payload-family corpus was not. The historical grader exhibited direct treatment leakage: treatment metadata gated the ATTACK_SUCCESS class, so fixed behavior could change class under treatment relabeling. A treatment-blind reconstruction corrects 58 historical ATTACK_SUCCESS or HIJACK_ATTEMPT labels to authorized benign completions while preserving three verified protected-data transfers and one separate unauthorized-forwarding case. The locked v2 census contains exactly zero ATTACK_SUCCESS records, while the forwarding case remains a HIJACK_ATTEMPT at a semantic boundary concerning objective completion. A dual-reviewer blinded concordance review of all 96 requests deemed structurally interpretable by locked v2 produced identical reviewer-consensus classes but differed from the locked codebook on four construct-boundary cases. We contribute a seven-link Integrity Chain and an executable, scope-bounded endpoint-integrity linter. The result is a campaign-bounded measurement audit, not a population attack-rate, model-ranking, defense-efficacy, or causal estimate.