PT-2026-49172

Name of the Vulnerable Software and Affected Versions python-utcp version 1.1.0 Description A server-side request forgery SSRF exists in the utcp-gql/utcp-websocket component of the universal-tool-calling-protocol. This issue allows a remote attacker to initiate a manipulation that results in the...

Universal Tool Calling Protocol 安全漏洞

Universal Tool Calling Protocol is an official Python implementation of the UTCP open-source protocol. Versions prior to 1.1.3 of Universal Tool Calling Protocol contained security vulnerabilities; these vulnerabilities stemmed from the prepareenvironment method passing complete environment...

Universal Tool Calling Protocol 代码问题漏洞

Universal Tool Calling Protocol is an official Python implementation of the UTCP open-source protocol. Versions prior to 1.1.3 of Universal Tool Calling Protocol had code vulnerabilities, which stemmed from inconsistent trust boundaries and could lead to man-in-the-middle server request forgery...

Universal Tool Calling Protocol 操作系统命令注入漏洞

Universal Tool Calling Protocol is an official Python implementation of the UTCP open-source protocol. Versions prior to 1.1.3 of Universal Tool Calling Protocol contained a vulnerability related to operating system command injection. This vulnerability stemmed from the substituteutcpargs method...

OrchJail: Jailbreaking Tool-Calling Text-To-Image Agents by Orchestration-Guided Fuzzing

Tool-calling text-to-image T2I agents can plan and execute multi-step tool chains to accomplish complex generation and editing queries. However, this capability introduces a new safety attack surface: harmful outputs may arise from tool orchestration, where individually benign steps combine into...

Your Agent Is More Brittle Than You Think: Uncovering Indirect Injection Vulnerabilities in Agentic LLMs

The rapid deployment of open-source frameworks has significantly advanced the development of modern multi-agent systems. However, expanded action spaces, including uncontrolled privilege exposure and hidden inter-system interactions, pose severe security challenges. Specifically, Indirect Prompt...

Trust Boundary Violation

Overview utcp is an Universal Tool Calling Protocol UTCP client library for Python Affected versions of this package are vulnerable to Trust Boundary Violation. Via the remote Manual Endpoint, the client retrieves a tool’s JSON specification, known as a Manual. An attacker can execute arbitrary...

Universal Tool Calling Protocol 安全漏洞

Universal Tool Calling Protocol is an official python implementation library for UTCP in the Universal Tool Calling Protocol open source. A security vulnerability exists in Universal Tool Calling Protocol that originates when a client obtains the JSON specification of a tool from a remote Manual...

Smart Tool Selection: Achieving 34-64% Token Savings with Spring AI's Dynamic Tool Discovery

As AI agents connect to more services—Slack, GitHub, Jira, MCP servers—tool libraries grow rapidly. A typical multi-server setup can easily have 50+ tools consuming 55,000+ tokens before any conversation starts. Worse, tool selection accuracy degrades when models face 30+ similarly-named tools. T...

Beyond JSON: Converting Spring AI Tool Response Formats to TOON, XML, CSV, YAML, ...

JSON is the go-to format for LLM tool responses, but recent discussions around alternative formats like TOON Token-Oriented Object Notation claim potential benefits in token efficiency and performance. While the debate continues—with critical analyses pointing to context-dependent results—the...

Exploiting Web Search Tools of AI Agents for Data Exfiltration

Large language models LLMs are now routinely used to autonomously execute complex tasks, from natural language processing to dynamic workflows like web searches. The usage of tool-calling and Retrieval Augmented Generation RAG allows LLMs to process and retrieve sensitive corporate data, amplifyi...

Mind the Gap: Evaluating Model- and Agentic-Level Vulnerabilities in LLMs with Action Graphs

As large language models transition to agentic systems, current safety evaluation frameworks face critical gaps in assessing deployment-specific risks. We introduce AgentSeer, an observability-based evaluation framework that decomposes agentic executions into granular action and component graphs,...

DoomArena: a Framework for Testing AI Agents against Evolving Security Threats

We present DoomArena, a security evaluation framework for AI agents. DoomArena is designed on three principles: 1 It is a plug-in framework and integrates easily into realistic agentic frameworks like BrowserGym for web agents and $τ$-bench for tool calling agents; 2 It is configurable and allows...