Chainlit’s MCP Endpoint Turned a Browser Request Into Server-Side Code Execution
Chainlit 2.12.0 fixes an unauthenticated MCP command-injection flaw and a blind SSRF path. MCP-enabled deployments need an urgent configuration-aware upgrade.
Rob Nidschelm
Tech tinkerer who spends too much time with home labs and Docker stacks. Into AI, Roman & Medieval history, and all things sci-fi. Avid traveller, reader and writer, with a passion for good food.
Gitea Patched the Hook. CISA Says Attackers Are Using It.
CISA says attackers are exploiting Gitea's diffpatch Git-hook flaw. Version 1.27.1 also fixes a separate unauthenticated file-read vulnerability.
Kaltura Trusted One URL. It Became File Read and Remote Code Execution.
Two unpatched flaws in Kaltura's HTML5 player expose local files and can give unauthenticated attackers code execution as the web-server user.
The Sandbox Failed. TRACE Tries to Prove What the Agent Actually Did.
Linux Foundation's TRACE project does not keep AI agents inside the sandbox. It creates hardware-attested records of what ran, under which policy and which tools were called.
The Agent Asked for Consent. Its Batch Tool Answered for the Operator.
AWS fixed a Strands Agents Tools flaw that let a crafted prompt use the batch tool to bypass human consent and execute Python on the agent host.
The Phone Was Stolen. An AI Voice Agent Asked the Owner to Unlock It.
AnonyMousKIT combines stolen-device data, phishing pages and low-cost AI voice agents to obtain the credentials needed to disable Activation Lock.
The Package Wasn’t the Payload. Its Trusted Mirror Was the Phishing Site.
Attackers used npm packages as storage for fake verification pages rendered through trusted mirrors, turning legitimate CDN domains into phishing infrastructure.
Alabama Is Testing Whether an AI Sandbox Failure Was Also a Consumer-Protection Failure.
Alabama has subpoenaed OpenAI over the Hugging Face agent intrusion, testing whether weak AI evaluation controls can also violate consumer-protection law.
LACMA Detected the Breach in July 2025. The Notice Came 13 Months Later.
LACMA says a July 2025 network breach exposed identity, financial and medical data. Public notification followed more than 13 months after detection.
SABnzbd’s Password Could Be Bypassed. A Download Could Plant the Code That Ran Next.
SABnzbd 5.1.2 fixes a critical unauthenticated route to code execution, a malicious-download path traversal that can poison trusted job state, and a separate authentication bypass. Internet-facing WebUIs are urgent, but the download-processing flaw can also affect local installations.