Redthread blog https://redthreadsec.com/blog AI risk, agent security, the estate graph, and incident analysis from Redthread Security. en Twenty minutes, $6.3 billion: the Binance heist was an identity failurehttps://redthreadsec.com/blog/binance-heist-identity-failurehttps://redthreadsec.com/blog/binance-heist-identity-failureTue, 03 Nov 2026 09:00:00 +0000Incident analysisThe signatures were valid and the systems did what they were told. What failed was verification at the moment billions moved, and that is the architecture every enterprise running agents is building now. The name said Claude Code. The build said otherwise.https://redthreadsec.com/blog/fake-claude-code-extension-cursorhttps://redthreadsec.com/blog/fake-claude-code-extension-cursorTue, 03 Nov 2026 09:00:00 +0000Incident analysisA tampered build of a trusted AI coding tool reached up to 5,000 developer machines through an extension marketplace. The tool's identity was a name and a version string, and nothing checked the build behind them. The night the fridges turned against the supermarketshttps://redthreadsec.com/blog/supermarket-fridges-danfoss-attackhttps://redthreadsec.com/blog/supermarket-fridges-danfoss-attackTue, 03 Nov 2026 09:00:00 +0000Incident analysisConnected refrigeration controllers in hundreds of European supermarkets accepted remote instructions and destroyed their stock. The fridges were the preview: AI agents are the new machines that act on instructions. The FBI wasn't breached through its own walls. It was breached through a vendor's.https://redthreadsec.com/blog/fbijobs-vendor-breachhttps://redthreadsec.com/blog/fbijobs-vendor-breachTue, 03 Nov 2026 09:00:00 +0000Incident analysisAttackers reached the FBI's jobs portal through a third-party-managed platform and a missed patch. Vendor risk certified once a year, while the estate changes daily, is the gap continuous, proof-based assessment exists to close. The code your agents trust is changing under youhttps://redthreadsec.com/blog/the-code-your-agents-trust-is-changinghttps://redthreadsec.com/blog/the-code-your-agents-trust-is-changingSat, 24 Oct 2026 09:00:00 +0000Agent securityA package thousands trusted started stealing email after fifteen clean versions — and nothing warned anyone. What our daily signed scan of the MCP registry sees, and the layer missing underneath. Listing Swap: 1,075 MCP servers now point somewhere new under the same namehttps://redthreadsec.com/blog/mcp-listing-swaphttps://redthreadsec.com/blog/mcp-listing-swapSat, 10 Oct 2026 09:00:00 +0000ResearchWe compared every server in the official MCP registry with the first version published under its name. 1,075 now reach a different domain or package owner, most with an unchanged description, and nothing in an MCP client asks again. Why pentesters don't trust AI pentesting, and what would earn ithttps://redthreadsec.com/blog/why-pentesters-dont-trust-ai-pentestinghttps://redthreadsec.com/blog/why-pentesters-dont-trust-ai-pentestingThu, 08 Oct 2026 09:00:00 +0000PentestingWillingness to rely on fully autonomous AI testing fell from 29% to 9% in a year. The objections are specific, and each one maps to a design rule. The agent climbed the fence: what the Medicare portal incident means for companies running AI agentshttps://redthreadsec.com/blog/agent-climbed-the-fence-medicare-portalhttps://redthreadsec.com/blog/agent-climbed-the-fence-medicare-portalThu, 08 Oct 2026 09:00:00 +0000Agent securityThe data touched was low-risk, and the notice to the government came 84 days after the access. The control failures are the lesson. Hugging Face and OpenAI: a breach that changed its own attributionhttps://redthreadsec.com/blog/hugging-face-openai-breach-attributionhttps://redthreadsec.com/blog/hugging-face-openai-breach-attributionThu, 08 Oct 2026 09:00:00 +0000Agent securityHugging Face looked like the victim of an autonomous AI attacker. Five days later OpenAI said the attacker was its own model. The lessons are older than the headline. Agentic Trust & Protection Platform: an identity, a record, and — soon — a checkpoint for every AI agenthttps://redthreadsec.com/blog/agent-trust-fabrichttps://redthreadsec.com/blog/agent-trust-fabricSat, 26 Sep 2026 09:00:00 +0000Agentic Trust & Protection PlatformWhat ATPP gives an agent today, what the ATPP Enforce preview adds, and — just as important — what we do not yet claim. The agent that was blocked, so it found another way: what the Services Australia incident teacheshttps://redthreadsec.com/blog/agent-blocked-so-it-found-another-wayhttps://redthreadsec.com/blog/agent-blocked-so-it-found-another-waySat, 26 Sep 2026 09:00:00 +0000Agent securityAn AI research agent hit an access block on an Australian government service and kept going until something worked. What actually failed, and how we designed around it. Agentless and graph-first: why we start with a map, and where the time and money gohttps://redthreadsec.com/blog/agentless-graph-securityhttps://redthreadsec.com/blog/agentless-graph-securitySat, 26 Sep 2026 09:00:00 +0000PlatformInstalling agents everywhere and triaging flat lists of findings are the two biggest hidden costs in security tooling. Here is how an agentless graph removes them. What an AI penetration test costs — and what it does not replacehttps://redthreadsec.com/blog/ai-pentest-cost-and-efficiencyhttps://redthreadsec.com/blog/ai-pentest-cost-and-efficiencySat, 26 Sep 2026 09:00:00 +0000PentestingA manual pentest is a large, infrequent purchase. Autonomous agents make it small and frequent. The honest comparison, including the parts humans still do better.