CVE-2026-55615

Langroid is a framework for building large-language-model-powered applications. Prior to version 0.65.5, Neo4jChatAgent passes LLM-generated Cypher queries straight to the Neo4j driver with no validation, no statement-type allowlist, and no opt-out gate. The query text is influenceable by prompt...

Directory Traversal

Overview litellm is a Library to easily interface with LLM API providers Affected versions of this package are vulnerable to Directory Traversal via path traversal in the skill archive extraction logic in SkillPromptInjectionHandler and SkillsSandboxExecutor. An authenticated user with access to...

CVE-2026-59807 Composio SDK < 0.2.32-beta.283 - Sensitive File Upload via tool-file-uploads.ts

Composio SDK before 0.2.32-beta.283 contains a path validation bypass vulnerability that allows attackers to read and exfiltrate sensitive files by exploiting a missing assertSafeFileUploadPath check in the readFileFromDisk function within tool-file-uploads.ts. Attackers can exploit prompt...

CVE-2026-59807 Composio SDK < 0.2.32-beta.283 - Sensitive File Upload via tool-file-uploads.ts

Composio SDK before 0.2.32-beta.283 contains a path validation bypass vulnerability that allows attackers to read and exfiltrate sensitive files by exploiting a missing assertSafeFileUploadPath check in the readFileFromDisk function within tool-file-uploads.ts. Attackers can exploit prompt...

CVE-2026-59807

CVE-2026-59807 affects Composio SDK prior to 0.2.32-beta.283. A path validation bypass in readFileFromDisk (tool-file-uploads.ts) can enable an attacker to read and exfiltrate sensitive files by manipulating file_uploadable parameters via prompt injection, referencing sensitive paths such as SSH ...

DRUPAL-CONTRIB-2026-076

The AI SEO/GEO Analyzer module generates SEO/GEO analysis reports by sending content of an entity including its comments to an LLM, then converts the model's Markdown response to HTML and stores it for display to privileged users. The generated HTML was rendered without passing through Drupal's...

New HalluSquatting Attack Could Trick AI Coding Assistants Into Installing Botnet Malware

AI coding assistants have a habit of making things up. Ask one to fetch a popular tool, and it will sometimes hand back a real-sounding name for a project that does not exist. New research, which its authors call HalluSquatting , turns that habit into an attack: work out the fake names an AI...

PT-2026-56665

Name of the Vulnerable Software and Affected Versions Drupal AI SEO/GEO Analyzer versions 0.0.0 through 1.1.3 Description Stored Cross-site Scripting XSS occurs when the module generates SEO/GEO analysis reports by sending entity content, including comments, to a Large Language Model LLM. The...

AI SEO/GEO Analyzer - Moderately critical - Cross-site Scripting - SA-CONTRIB-2026-076

The AI SEO/GEO Analyzer module generates SEO/GEO analysis reports by sending content of an entity including its comments to an LLM, then converts the model's Markdown response to HTML and stores it for display to privileged users. The generated HTML was rendered without passing through Drupal's...

PYSEC-2026-1741 Open WebUI vulnerable to Stored DOM XSS via prompts when 'Insert Prompt as Rich Text' is enabled resulting in ATO/RCE

Summary The functionality that inserts custom prompts into the chat window is vulnerable to DOM XSS when 'Insert Prompt as Rich Text' is enabled, since the prompt body is assigned to the DOM sink .innerHtml without sanitisation. Any user with permissions to create prompts can abuse this to plant ...

PYSEC-2026-1555 LlamaIndex vulnerable to Creation of Temporary File in Directory with Insecure Permissions

A vulnerability in the defaultjsonalyzer function of the JSONalyzeQueryEngine in the run-llama/llamaindex repository allows for SQL injection via prompt injection. This can lead to arbitrary file creation and Denial-of-Service DoS attacks. The vulnerability affects the latest version and is fixed...

PYSEC-2026-1507 Langchain SQL Injection vulnerability

A vulnerability in the GraphCypherQAChain class of langchain-ai/langchain version 0.2.5 allows for SQL injection through prompt injection. This vulnerability can lead to unauthorized data manipulation, data exfiltration, denial of service DoS by deleting all data, breaches in multi-tenant securit...

Open WebUI Stored Cross-Site Scripting Vulnerability

Attackers can craft a malicious prompt that coerces the language model into executing arbitrary JavaScript in the context of the web page...

PYSEC-2026-1508 Langchain Server-Side Request Forgery vulnerability

In Langchain before 0.0.329, prompt injection allows an attacker to force the service to retrieve data from an arbitrary URL, essentially providing SSRF and potentially injecting content into downstream tasks...

Langchain Server-Side Request Forgery vulnerability

In Langchain before 0.0.329, prompt injection allows an attacker to force the service to retrieve data from an arbitrary URL, essentially providing SSRF and potentially injecting content into downstream tasks...

PYSEC-2026-1757 pandasai vulnerable to prompt injection

An issue in Gaberiele Venturi pandasai v.0.8.0 and before allows a remote attacker to execute arbitrary code via a crafted request to the prompt function...

Langroid: Neo4jChatAgent executes LLM-generated Cypher without validation (prompt-to-Cypher injection; config-conditional RCE), mirroring the SQLChatAgent bug fixed in CVE-2026-25879

Neo4jChatAgent passes LLM-generated Cypher queries straight to the Neo4j driver with no validation, no statement-type allowlist, and no opt-out gate. The query text is influenceable by prompt injection direct user input or indirect content the agent reads back via RAG, so an attacker who can...

GHSA-2PQ5-3Q89-J7CC Langroid: Neo4jChatAgent executes LLM-generated Cypher without validation (prompt-to-Cypher injection; config-conditional RCE), mirroring the SQLChatAgent bug fixed in CVE-2026-25879

Neo4jChatAgent passes LLM-generated Cypher queries straight to the Neo4j driver with no validation, no statement-type allowlist, and no opt-out gate. The query text is influenceable by prompt injection direct user input or indirect content the agent reads back via RAG, so an attacker who can...

GHSA-Q9P7-WQXG-MRHC Langroid: Sandbox Escape to Remote Code Execution via Incomplete `eval()` Mitigation in TableChatAgent

Advisory Details Title: Sandbox Escape to Remote Code Execution via Incomplete eval Mitigation in TableChatAgent Description: Summary Langroid is vulnerable to a critical Sandbox Escape leading to Remote Code Execution RCE in its TableChatAgent and VectorStore capabilities. When these agents...

Langroid: Sandbox Escape to Remote Code Execution via Incomplete `eval()` Mitigation in TableChatAgent

Advisory Details Title: Sandbox Escape to Remote Code Execution via Incomplete eval Mitigation in TableChatAgent Description: Summary Langroid is vulnerable to a critical Sandbox Escape leading to Remote Code Execution RCE in its TableChatAgent and VectorStore capabilities. When these agents...