Wiz presents three investigated intrusions involving an exposed Flowise service, a Java operations gateway and an MCP-facing GraphQL backend. It proposes seven behavioral fingerprints for agentic activity, combining tool adaptation, natural-language artifacts and execution patterns, while describing humans as setting objectives and guardrails.
Tools & MCP
Risks in the tools, MCP servers, plugins, and connectors agents use.
Tools, MCP, and the agent supply chain
How the tools and packages agents rely on became an attack path, from poisoned tool descriptions to malicious MCP servers in public registries.
- Invariant Labs discloses MCP tool poisoning, rug pull and shadowing attack classesThe tool-poisoning disclosure.
- Invariant Labs shows GitHub MCP agents can be steered by a public issue to leak private repo dataA broadly scoped token turns one injected issue into private data exposure.
- Text in an MCP tool's description can steer an agent's other actions, including leaking local secrets.What the evidence shows about tool descriptions.
- Approving an agent's tools or configuration once is not enough, because they can change after approval or be changed by the agent itself.Why approving a tool once is not enough.
- Malicious postmark-mcp npm package quietly copied every sent email to an outside addressA trojanized MCP server used against real users.
- Malicious or compromised agent extensions, MCP servers, and skills have been published to public registries and used against real users.The pattern across incidents.
Use the arrow keys to move between records, Home and End to jump to the first and last, and Enter to select one.
Select a mark to read the record. Mark size shows editorial significance. Hollow marks are dated to the month. Era bands are editorial labels.
Records in view
41 records · newest firstForever Security reports embedded-browser weaknesses in Cursor, GitHub Copilot, Claude Desktop, Codex and VS Code, with different impacts including host execution, address spoofing and local-file access. The researchers describe controlled demonstrations and product-specific fixes or unresolved reports; their account does not establish exploitation against users.
JFrog reports 42 armed third-party GitHub repositories across 11 legitimate owners linked to the testmetest npm campaign. Recovered artifacts included malicious publishers and execution hooks for Claude, Gemini, Cursor and VS Code; repository counts do not establish the number of infected developer machines or downstream executions.
V12 reports 25 vulnerabilities in x402 code, detailing three rated critical: payment verification without a transfer, signing for an unrelated website, and automated payment authorization without spending limits. It says all three critical issues were fixed by 2026-08-20; the report describes local tests and possible impacts, not observed wallet theft.
Johann Rehberger (Embrace The Red) reports that a file-path check meant to stop Copilot in Visual Studio and SQL Server Management Studio from writing to its own protected configuration could be bypassed, letting the agent change its settings, escape its containment restrictions and run arbitrary commands. He says he also showed that a third party could trigger this through indirect prompt injection. Microsoft classified the issue Important (remote code execution) and, the post says, shipped the fix on 2026-10-13. He says he found it while validating fixes for earlier SQL Copilot security issues; these are the researcher's claims, not a Microsoft advisory.
AIR Security reports a controlled demonstration of data leaving Vercel Sandbox through an allowed external service, despite a domain allowlist blocking a direct connection to the researchers’ receiver. The external service made an onward request while processing agent-controlled certificate data. This is an indirect egress finding, not a demonstrated VM escape or an observed rogue-agent incident; AIR says Vercel classified the report as outside its bounty scope because the technique depends on a server Vercel does not operate.
METR reports a proof of concept in which untrusted agent output could change what a reviewer saw or downloaded through the Inspect transcript viewer. Under METR’s infrastructure, the real trajectory would still exist in its database, and METR says it has not observed agents exploiting this flaw in its evaluations. Meridian Labs patched the vulnerability within a day of reporting; METR argues that observability tools should be treated as security-critical infrastructure.
A GitHub security advisory in AWS Labs' Loom repository, published 2026-10-02, reports that in Loom, an open-source platform for building, deploying and managing AI agents, the backend treated every request as a super-administrator when no identity provider was configured, including requests with no credentials. The advisory rates it Critical, CVSS 10.0, and says it affects versions before 1.6.1, fixed in 1.6.1, released 2026-08-04. It says a network client that could reach the backend could read, change or delete managed resources, including agents, memories, authorizer configuration and credentials, and invoke any agent.
GitLab, as the CVE numbering authority, published CVE-2026-90970 on 2026-10-02: in the AI Gateway component, an authenticated user with Duo Agent Platform access could escape the prompt template sandbox through a specially crafted flow configuration and run arbitrary commands on the gateway. The CVE record rates it critical (CVSS 3.1 base score 9.9) and lists fixes in versions 19.2.4, 19.3.2 and 19.4.1. The record also carries a CISA assessment, added the same day, that lists exploitation as none; The Hacker News adds that only organizations hosting their own gateway need to act.
Salt Labs reports a researcher-controlled email injection that caused Manus to run untrusted code when asked to read an inbox, before displaying a security warning. The researchers describe access to the user’s sandbox and connected-account tokens, including Gmail and potentially other connected services. Salt says the reported flaw has been resolved and is no longer exploitable; the post does not establish an attack campaign against real users.
Researchers from the Chinese Academy of Sciences and Worcester Polytechnic Institute introduce TrustProbe to trace installed skill content into security-sensitive operations and verify observable effects. Using DeepSeek-V4-Flash across eleven open-source agents, they report 104 verified source-to-sink vulnerabilities in permissive non-interactive configurations, with a subset remaining exploitable under stricter approval settings. Their real-skill experiment measures exposure to vulnerable execution paths, rather than the prevalence of malicious skills or attacks on real users.
Two Huawei Research Zurich authors present Pretext, a white-box LLM attacker that iteratively writes agent skills to evade a scanner pairing static rules with LLM-based semantic analysis, using NVIDIA's open-source SkillSpector as the target. Their abstract reports up to 97% attack success against a frozen scanner and up to 77% against one that also learns, across three open-source model stacks; the body shows attack success falling over generations against the learning scanner (final-generation values are not tabulated) and finds that the learning scanner became more conservative with high false-positive rates. Commercial scanners were not tested.
Johann Rehberger (Embrace The Red) writes up a BlueHat Asia 2026 talk on Microsoft's Copilot in SQL Server Management Studio, reporting that the assistant runs database actions with the privileges of whoever is connected and that its "read-only" mode rested on a prompt instruction and a pattern-matching filter rather than a permission. He says this let injected instructions, including ones planted in database content or metadata, drive arbitrary database actions, data exfiltration to a third-party server and, in his final demonstration, a lower-privileged database owner becoming sysadmin. Microsoft rated the resulting CVE critical; the post urges readers to update their installations and gives no patch version or date.
Google Threat Intelligence Group (GTIG) reports that monthly CVE disclosures rose from 5,045 in January 2026 to 10,477 in July and 10,740 in August, and that exploited vulnerabilities averaged 18 per month from January to August 2026 against 10.5 per month in 2025. GTIG says zero-day exploitation rose only marginally and suggests rapid weaponization of n-days drove most of the growth, and that vulnerabilities it identified as likely AI-discovered skew toward Medium risk ratings and remote code execution. It also counts 2,076 cumulative AI-related CVEs (January 2025 to August 2026) and reports in-the-wild exploitation of a handful, including a flaw in LiteLLM's MCP server preview endpoints.
Li, He, Dai and Xiao (arXiv v1, 29 September 2026) propose ToolFence, an inference-time defense that compiles the authenticated user request into typed capabilities with provenance constraints on authority-sensitive arguments, enforces them with a deterministic monitor, and asks an LLM judge to grant new capability shapes rather than judge each call. On AgentDojo with Qwen3-max the authors report overall attack success of 0.20% against 21.20% undefended, with clean utility 38.90% against 42.70% undefended; a cross-session cache of approved shapes cut judge calls per task from 1.84 to 1.05 in their ablation. The attacks are six injection strategies averaged; the authors report no adaptive attacks against ToolFence and residual failures inside authorized data flows.
Independent researcher Animesh Shaw (arXiv v1, 26 September 2026) audits one unnamed indirect-prompt-injection benchmark and describes four defect classes: payloads silently not delivered, attack success scored by tool identity rather than arguments, false rejection conflated with model incapacity, and no audit trail. Re-scoring recorded traces from the author's own runs, the tool-identity scorer gives 21.7% attack success against 1.2% at argument level, and a model reported at 62.8% by the audited harness scores 0% on the corrected one, a gap the author attributes in substantial part to undelivered payloads and tool-identity scoring. The author also releases a harness that rejects measurement-invalid scenarios before running, and says the results use static attacks in a five-tool simulation and are not robustness claims.
The Australian Signals Directorate's ACSC published guidance on agentic AI harnesses, the software layer that connects a model with organisational data, tools and systems and manages context, memory, tool access and execution privileges. According to coverage, it says some risks, including prompt injection, cannot be addressed within the model alone, that no harness is inherently secure, and recommends least privilege, human oversight for high-impact actions, audit logging and validating agent outputs before execution.
An author project page describes DeltaCert-Agent, which maps configuration changes in tool-using LLM agents to affected security claims and reruns only scoped tests plus sentinel checks, escalating to full recertification when impact cannot be bounded. The author reports 75.02% regression-detection recall versus 55.01% for equal-budget random selection while running 61.35% fewer tests, using four small locally hosted models.
The NSA's Artificial Intelligence Security Center released a cybersecurity information sheet on the Model Context Protocol, warning that adoption has outpaced safeguards. It recommends vetting MCP tools, least-privilege access and isolation, validating outputs where one model's output feeds another, and detailed logging integrated with security monitoring, and it lists poor approval workflows among the risks.
Microsoft's Defender security research team described two vulnerabilities in its open-source Semantic Kernel agent framework in which attacker-influenced prompts could reach code execution on the host. One involved unsafe string interpolation in an in-memory vector store filter; the other exposed a code-execution plugin function that allowed arbitrary file writes outside its sandbox. Both were fixed in updated SDK releases.
Pluto Security found that nginx-ui's MCP integration left a message endpoint effectively unauthenticated under default settings, letting anyone reach its administrative MCP tools and take over the web server. The flaw (CVSS 9.8) was fixed in version 2.3.4 on 2026-03-15, and Recorded Future listed it among actively exploited vulnerabilities in March 2026.
OX Security reports that MCP's STDIO transport turns configuration into OS command execution, and that frameworks exposing that configuration to users, networks or prompt injection inherited remote code execution, with 12+ CVEs across projects such as LangFlow, LiteLLM, Flowise and Windsurf. The Hacker News reports Anthropic characterized the protocol behavior as expected and did not change the reference design.
Check Point Research found that a cloned repository's Claude Code configuration could run hooks, start MCP servers before the user approved them, and redirect API traffic so the user's Anthropic API key was sent to an attacker (CVE-2025-59536, CVE-2026-21852). Anthropic fixed the issues between August and December 2025 by deferring execution and API calls until after the trust dialog.
CAISI, with NIST's Information Technology Laboratory, launched an initiative to support industry-led standards and open protocols for AI agents and research on agent security and identity. Announced deliverables included the agent security RFI, an AI agent identity and authorization concept paper with comments due April 2, 2026, and sector listening sessions from April 2026.
Cyata found path-validation and argument-injection flaws in Anthropic's mcp-server-git (CVE-2025-68143, -68144, -68145) that, chained with the Filesystem MCP server, allowed file access and code execution by anyone able to influence what the assistant reads. Anthropic removed the git_init tool and added path validation in fixed releases.
The OWASP GenAI Security Project released its Top 10 for Agentic Applications, a list of ten risk categories specific to agents that plan, hold memory, call tools and act with delegated authority. The release came with an updated Agentic Threats and Mitigations taxonomy (v1.1) and a capture-the-flag practice platform.
MITRE ATLAS version 5.0.0 added a set of techniques for attacks on AI agents, including agent context poisoning of memory and threads, modifying agent configuration, credential theft from agent configuration, and exfiltration via agent tool invocation, and renamed LLM Plugin Compromise to AI Agent Tool Invocation. Version 5.1.0 (November 6, 2025) added agent-specific mitigations such as tool permission configuration and human-in-the-loop for agent actions.
A package impersonating a Postmark email MCP server was published to npm and, after 15 clean versions, version 1.0.16 (2025-09-17) added code that blind-copied all emails sent through it to the publisher. Postmark stated it had never published an MCP server on npm; Koi Security found it, and the package was deleted after about 1,643 downloads.
Wang and colleagues build MCPTox from 45 real MCP servers and 353 authentic tools, generating 1,312 malicious test cases across 10 risk categories. Across 20 LLM agents the highest attack success rate was 72.8% (o1-mini), and refusals were rare, with the highest refusal rate under 3% (Claude 3.7 Sonnet).
Zenity Labs presented at Black Hat USA 2025 a set of zero- and one-click prompt injection chains, including a shared document causing ChatGPT Connectors to search a victim's Google Drive for API keys and leak them through image rendering, and a poisoned email steering a Copilot Studio agent to disclose CRM data. CSO Online reports that OpenAI and Microsoft deployed fixes for the specific demonstrated techniques.
SafeBreach researchers showed that instructions in calendar invite titles or email subjects, retrieved when a user asked Gemini about their schedule, could trigger tool misuse such as deleting events, opening URLs, exfiltrating email subjects, and controlling connected smart-home devices. They disclosed to Google in February 2025, and Google deployed layered mitigations including confirmations, URL handling policies and injection classifiers.
Check Point Research found that Cursor bound MCP approval to a configuration's name rather than its contents, so a collaborator with repository write access could swap an approved harmless command for a malicious one that ran on each project open. Cursor 1.3, released 2025-07-29, prompts for approval on any MCP configuration change.
Cursor's advisory states that the agent could create new workspace dotfiles without approval, so injected instructions arriving via an external MCP source could write an MCP configuration that launched attacker commands. Aim Security researchers reported it; it is rated CVSS 8.5 and fixed in Cursor 1.3.9.
JFrog reported CVE-2025-6514 (CVSS 9.6) in mcp-remote, a proxy used by MCP clients to reach remote servers, where a malicious server could supply a crafted OAuth authorization URL that led to command execution on the client machine. Versions 0.0.5 to 0.1.15 are affected and 0.1.16 fixes the issue.
General Analysis demonstrated a Cursor agent connected to Supabase MCP with a service-role key, which bypasses row-level security, following instructions in a customer support ticket to read a secrets table and write the contents back into the attacker-visible ticket. Supabase later responded that agents should not be connected to production data and described guardrails that reduced but did not eliminate risk.
The 2025-06-18 revision of the Model Context Protocol specification classifies MCP servers as OAuth resource servers with protected resource metadata, and requires clients to implement RFC 8707 resource indicators so malicious servers cannot obtain tokens meant for others. It also clarifies authorization security considerations and adds a security best practices page.
Authors from Invariant Labs, IBM, ETH Zurich, Swisscom, Google, Microsoft and others propose six design patterns that limit what injected text can do, such as action-selector, plan-then-execute, dual LLM, LLM map-reduce, code-then-execute and context minimization. They discuss utility and security trade-offs and illustrate the patterns with ten application case studies.
Invariant Labs demonstrated that a malicious issue in a public repository could lead an agent using the GitHub MCP server to read the user's private repositories and publish the data in a public pull request. The firm tested with Claude 4 Opus and argues there is no server-side patch because the flaw lies in agent permissions, recommending per-session repository scoping and runtime monitoring.
Invariant Labs describes tool poisoning, in which instructions hidden in an MCP tool's description are visible to the model but not to the user, and shows proof-of-concept exfiltration of local files through an MCP client. It also describes rug pulls, where a server changes tool descriptions after approval, and shadowing, where one server's descriptions alter how the agent uses another server's tools. Recommended mitigations include showing full tool descriptions, pinning tool versions with checksums, and cross-server isolation.
Hou, Zhao, Wang and Wang survey MCP's architecture, industry adoption and server lifecycle. The first version (March 2025) split the lifecycle into creation, operation and update phases and discussed security risks in each. A revision in October 2025 expanded this to four phases with 16 activities and a threat taxonomy of four attacker types and 16 threat scenarios, with case studies and per-phase safeguards.