Skip to main content
Mallory
← Back to News
ai-platform-security

Copycat MCP Tools Redirect AI Agents to Attacker-Controlled Providers in Controlled Tests

Emerging Technology Risk · 3 sources · Updated 17h ago · First seen Nov 17, 2026

Copycat MCP Tools Redirect AI Agents to Attacker-Controlled Providers in Controlled Tests

Researchers demonstrated tool-provider misbinding in AI agents using the Model Context Protocol (MCP): requests intended for a legitimate provider were routed to an attacker-controlled tool advertising copycat metadata. In controlled evaluations across five LLM-based clients, matching tool names, copied descriptions and compatible schemas caused GPT-5.4, Qwen3.7 Max and DeepSeek V3 to select the attacker’s tool in all 50 trials each, versus 84% for Gemini 2.5 Pro and 6% for Claude Sonnet 5. The attack required the attacker’s MCP server to be connected to the agent’s tool environment, but did not require compromising existing infrastructure or embedding malicious instructions.

The strongest failures occurred when tool-name collisions, task-aligned descriptions and compatible schemas reinforced one another. Descriptions asserting authority, paired with compatible schemas, also caused wrong-provider execution or unresolved ambiguity without matching names, depending on the model. Claude primarily requested clarification, preventing automatic attacker-tool execution in those trials without establishing provider identity. The findings highlight the risk of treating tool metadata as proof of provider identity. The authors proposed provider-bound identities and signed tool manifests, but did not evaluate those defenses; the single-task preprint has not been independently replicated.

New surface, new adversariesTrack how attackers are adapting to this technology.
Track this risk
Event timeline

How this story unfolded

Events from the most recent confirmed update back to the earliest known activity.

3 events · 2025 – Nov 16, 2026
  1. Reported Nov 18, 20264d ago

    Cloud Security Alliance whitepaper links tool squatting to unverified tool identity

    The Cloud Security Alliance’s AI Safety Initiative published a whitepaper linking tool squatting to missing cryptographic verification of tool-name identity. It described the Enhanced Tool Definition Interface as a signed-definition research proposal not incorporated into MCP.

    GPT-5.4, Qwen3.7 Max and DeepSeek V3 Picked Copycat MCP Server Tool Over Real One in All 50 Trials
  2. Nov 16, 20266d ago

    Researchers submit MCP tool-provider misbinding study covering five AI clients

    Zeynab Anbiaee, Sajjad Dadkhah and Ali A. Ghorbani submitted a preprint showing that attacker-controlled MCP tool metadata can redirect requests intended for legitimate providers without compromising existing infrastructure. With identical tool names, copied descriptions and compatible schemas, GPT-5.4, Qwen3.7 Max and DeepSeek V3 selected the attacker’s tool in all 50 trials each, compared with 84% for Gemini 2.5 Pro and 6% for Claude Sonnet 5.

    [2611.21577] When Tool Names Are Not Identities: Measuring Provider Misbinding in MCP-Based AI Agents
  3. 2025

    University of Delaware study finds tool-routing confusion in Cursor

    A 2025 study by Xiaofan Li and Xing Gao found that the tested Cursor version invoked the first-listed tool regardless of the model’s selection; the behavior was not observed in Windsurf, Claude Desktop or Cline. The researchers also examined 44,499 tools and found the name "add" used 379 times across servers.

    GPT-5.4, Qwen3.7 Max and DeepSeek V3 Picked Copycat MCP Server Tool Over Real One in All 50 Trials
Linked entities

Related entities

Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.

Open in app →

Affected products

1 linked
Deepseek-V3

Organizations

2 linked
DeepseekAnthropic
Do you run this?Follow how adversaries are adapting to this technology, and where it touches your stack today.
Track This Risk
Mallory platform

See the full picture, correlated to your attack surface.

This page covers what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which threat actors are using it right now, and what to deploy next.

Exposure mappingMap indicators from this story to your assets and identify affected systems in minutes.
Threat actor evidenceEvery observed campaign, victim, and pivot linked to actors named in this story.
Associated malwareMalware, exploits, and IOCs connected to the activity described here.
Detection signaturesYARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Scheduled alertsGet matching new stories delivered to your team as they break, not the next morning.
AI threadsAsk questions about this story and take action on the answers.