Cascading Failures

Researchers identify 33 protocol-level vulnerabilities across three leading agentic commerce platforms, achieving a 100% attack success rate independent of the AI model used, with three vulnerabilities chaining into a full payment hijack. This is a research paper (not an active exploit in the wild) demonstrating that agent-to-commerce-service protocols, not model behavior, are the primary structural risk in agentic payment systems.

Updated Jul 27, 2026

A threat actor reportedly leveraged the open-source Hermes AI agent running in an unattended 'YOLO' (no human confirmation) mode to automate post-exploitation actions during a breach of Thailand's Ministry of Finance. This represents real-world weaponization of agentic AI frameworks to accelerate attacker operations rather than a novel vulnerability in the agent itself, but it highlights the risk of autonomous, unsupervised agents executing tools with elevated privileges against production/government systems.

Updated Jul 25, 2026

A logic flaw in n8n's workflow automation platform allows an authenticated user to craft a request that associates their newly created workflow with a folder belonging to a different, unauthorized project. The workflow itself stays private to the attacker's project and no cross-project data is exposed, limiting the impact to a database-level integrity issue in folder structure. This is a traditional web application IDOR/authorization bug rather than an AI-agent-specific exploit, though n8n is increasingly used to orchestrate AI agent workflows.

Updated Jul 23, 2026

An authenticated n8n user can repeatedly abuse the data-table file upload endpoint to bypass per-request quota checks, causing temporary files to accumulate on disk faster than cleanup can remove them. This can exhaust host disk space, leading to a denial-of-service condition affecting the n8n instance and potentially other services on the same host. This is a genuine, moderate-severity resource exhaustion flaw rather than a novel agentic-AI attack, but it is directly relevant to n8n's growing role as an orchestration layer for AI agent workflows.

Updated Jul 23, 2026

A REST endpoint in the n8n workflow automation platform is registered before authentication middleware is applied, letting any unauthenticated network caller cancel another user's active test webhook if they know the workflow ID. The impact is limited to disrupting in-progress test sessions; production webhooks and stored data are unaffected. This is a low-complexity availability/harassment issue rather than a data breach or compromise vector.

Updated Jul 23, 2026

LiteLLM's Skills feature fails to validate file paths when extracting uploaded ZIP archives, allowing an authenticated user with access to the Skills API to write files outside the intended directory. This 'zip-slip' style vulnerability could lead to arbitrary file write and potentially remote code execution depending on server configuration. The issue is fixed in version 1.83.7-stable.

Updated Jul 23, 2026

LiteLLM's /health/test_connection endpoint could be used by a privileged proxy administrator or authorized caller to read arbitrary local files by supplying an oidc/file/ reference in litellm_params. Exploitation requires existing privileged access to the proxy's model-testing functionality, so the practical impact is limited to a defense-in-depth weakness rather than a cross-tenant or unauthenticated bypass. The issue is fixed in version 1.83.10-stable.

Updated Jul 23, 2026

Network-AI's MCP SSE server ships with an empty default authentication secret, meaning every request is treated as authorized regardless of whether an Authorization header is present. Combined with a wildcard CORS policy on all responses, this lets a malicious webpage silently invoke any of the 22 exposed MCP tools on a victim's local server, including spawning agents and writing to shared blackboard state. This is a classic drive-by CSRF-style attack against a locally running privileged agent orchestrator.

Updated Jul 20, 2026 · CVSS 7.6

Hugging Face disclosed that an autonomous AI agent was used to breach its production infrastructure, resulting in unauthorized access to internal datasets and credentials. The incident is notable because the attack vector was an AI agent operating with some degree of autonomy rather than a purely manual intrusion, highlighting real-world risk of agentic systems being weaponized against AI platform infrastructure. Details remain limited, as the source article is truncated and lacks technical specifics on the agent's tooling or exploitation method.

Updated Jul 20, 2026

NadMesh is a Go-based botnet that scans the internet for unauthenticated, publicly exposed AI infrastructure (ComfyUI, Ollama, n8n, Open WebUI, Langflow, Gradio) and abuses them to harvest cloud credentials, including a claimed 3,811 unique AWS keys, as well as Kubernetes tokens. This is not an attack on agent reasoning or protocols themselves, but exploitation of poor deployment hygiene around AI tooling that teams stand up quickly without adequate authentication or network controls. Severity is high because successful compromise yields direct cloud/infrastructure takeover, not just AI misuse.

Updated Jul 19, 2026

IBM Langflow OSS versions 1.0.0-1.10.0 allow unauthenticated attackers to register new user accounts that, under the common NEW_USER_IS_ACTIVE=true configuration, are immediately active without admin approval. This lets attackers authenticate and reach code-execution endpoints, effectively bypassing intended access controls even when AUTO_LOGIN is disabled. Given the critical CVSS score of 9.8, this is a genuine and severe vulnerability requiring urgent patching.

Updated Jul 18, 2026 · CVSS 9.8

This is a follow-up disclosure from Embrace The Red detailing how an LLM agent that outputs untrusted content into a macOS Terminal could be leveraged to trigger DNS requests via crafted ANSI escape sequences, effectively exfiltrating data through DNS lookups. Apple has since fixed the underlying macOS Terminal behavior, so this report documents a resolved vulnerability rather than an active ongoing threat. Severity is moderate given the fix is already deployed, but the technique remains relevant for agents/tools still rendering raw terminal output without sanitization.

Updated Jul 17, 2026

AWS Bedrock AgentCore Python SDK versions 1.4.8 and 1.5.0 logged raw user prompts and full agent responses into OpenTelemetry span attributes without any filtering or masking, which then flowed into customer CloudWatch aws/spans log groups. Any principal with read access to those logs could view sensitive user content, including secrets or PII that may have been part of prompts or agent outputs. This is a data exposure issue rather than an active exploit, and requires local/authenticated access to the customer's own logging infrastructure.

Updated Jul 16, 2026 · CVSS 5.7

xAI's Grok Build CLI coding agent was found to upload the entire working directory (and in one reported case, a user's home directory including SSH keys and password manager databases) to xAI's Google Cloud storage without clear user consent. xAI disabled the feature, deleted retained data, and open-sourced the codebase in response to backlash, but this represents a serious real-world data exfiltration incident caused by an overly broad and opaque agent tool/data-retention design rather than a targeted attack.

Updated Jul 16, 2026

A Drupal contributed module named 'AI Agents' has an access control flaw that allows forceful browsing, meaning unauthorized users may reach pages or functionality they shouldn't have access to. This is a standard web application authorization bug rather than a novel AI agent security issue, and no evidence in the data suggests prompt injection, tool poisoning, or agent-to-agent exploitation. Severity is moderate given the lack of CVSS score and limited detail on what functionality is exposed.

Updated Jul 11, 2026

A Drupal contributed module called 'AI Agents' has a missing authorization vulnerability allowing forceful browsing, meaning certain pages or endpoints may be accessible without proper permission checks. This is a standard web application access-control flaw in a CMS module that happens to relate to AI agent functionality, rather than a novel AI-specific attack technique. Severity is moderate since exploitation depends on the module's deployment and what functionality is exposed without authorization.

Updated Jul 11, 2026

LiteLLM's MCP Streamable HTTP endpoint prior to version 1.84.0 could be tricked into skipping key validation by sending a fabricated Authorization header, causing the server to fall back to an empty, effectively unauthenticated user context. This let attackers reach MCP tooling and any downstream tools/agents exposed through the proxy without holding a valid LiteLLM API key. The vulnerability is fixed in 1.84.0.

Updated Jul 9, 2026

A low-severity vulnerability was identified in LangGraph's Task Result Cache where the internal _freeze function uses a weak hash for generating default cache keys. Exploitation requires high attack complexity and remote access, with a CVSS score of 3.1, making practical exploitation difficult. A fix is pending via an open pull request.

Updated Jul 5, 2026 · CVSS 3.1