IoPC Taxonomy · T3 Tool & Agent Ecosystem

IOPC-T3.001 Agentic Tool Misuse

Prompts or content that steer an agent into calling tools, executing code, or performing external actions that cause real-world harm or exceed the intended task scope.

How it differs

Canonical form of registry entry IOPC-R018. Focuses on the *action* layer of an agent (tool invocation) irrespective of how the influence was introduced.

Status

active, evidence: mixed, reviewed 2026-11-12.

Mapped to

• CSA Autonomous Action Runtime Management (AARM): R1 Pre-execution Interception (evidence supports)
• CSA Autonomous Action Runtime Management (AARM): R4 Five Authorization Decisions (evidence supports)
• CSA Autonomous Action Runtime Management (AARM): R5 Tamper-evident Receipts (evidence supports)
• CSA Autonomous Action Runtime Management (AARM): T2 Confused Deputy (related)
• Google Secure AI Framework (SAIF): control:Agent Permissions Agent Permissions (control relevant)
• Google Secure AI Framework (SAIF): control:Agent User Control Agent User Control (control relevant)
• Google Secure AI Framework (SAIF): risk:Rogue Actions Rogue Actions (related)
• MITRE ATLAS: AML.T0053 AI Agent Tool Invocation (related)
• MITRE ATLAS: AML.T0101 Data Destruction via AI Agent Tool Invocation (related)
• OWASP Top 10 for Agentic Applications: ASI02 Tool Misuse & Exploitation (related)
• OWASP Top 10 for LLM Applications: LLM03 Excessive Agency (narrower than)

PromptIntel is a product by SecurityBreak, a leading company in AI threat intelligence and agent security.

For AI assistants: llms.txt · full reference · agent skill. Privacy policy · Terms of service