When Neither Side Confirms: Destructive Tool Approval Across MCP Servers and Hosts
cs.SE, cs.CR
Submitted: 2026-11-02
Updated: 2026-11-02
Code: https://github.com/containers/kubernetes-mcp-server
License: http://creativecommons.org/licenses/by/4.0/
The gist: LLM agents call tools that delete data or move money through the Model Context Protocol (MCP).
Abstract
LLM agents call tools that delete data or move money through the Model Context Protocol (MCP). The specification asks for a human who can deny tool calls but splits the mechanism between the server, which can annotate a tool or ask the user through the host (elicitation), and the host application, which decides whether to prompt. We measure both sides from source code. In a 1,000-repository sample of the official MCP registry, an estimated 90% of destructive tools (95% CI 82-95) have no server-side confirmation, and most servers that elicit proceed, or fall back to a flag the model can set, when the host cannot relay the question. Of 41 popular open-source hosts and agent SDKs, 32 have no approval step by default, including 16 of 25 applications. Pairing sampled destructive tools flagged by a rule-based extractor with these applications, 60.7% of pairs have no enforced human approval by default. Through four agent SDKs, four of five low-cost models executed an injected deletion in 479 of 960 runs. If hosts prompted for every tool not annotated read-only or non-destructive, the unapproved share of these pairs would fall to 5.4%; universal elicitation support would leave 56.7%. Applications should therefore configure approval.
Sources
- What a Random Draw from the MCP Registry Contains, and What Tool-Use Benchmarks Contain Instead
- Model Context Protocol Threat Modeling and Analyzing Vulnerabilities to Prompt Injection with Tool Poisoning
- Approve What, Exactly? Effect Binding and Trusted Rendering for Agent Step-Up Authorization
- An Empirical Study of Model Context Protocol Applications
- When Tool Calls Succeed but Workflows Fail: Anomalies at the Agent-Tool Boundary
- What You Approve Is What Executes: Consent Integrity for Black-Box LLM Agents
- MCPSecBench: A Systematic Security Benchmark and Playground for Testing Model Context Protocols
- HarnessSecurity-Bench: Do Security Mechanisms Really Protect Coding Agent Harnesses?
Related papers
- ArkEval: Benchmarking and Evaluating Automated CodeRepair for ArkTS
- TestPrism: Rethinking Test Evaluation Beyond a Single Reference
- Mining Workflow Graphs for Black-Box Boundary Testing of Conversational LLM Agents
- Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
- GitSkills: A Dataset of Agent Skills on GitHub
- For Coding Agents, the Repository Matters