Project
The AI output-inspection product whose SlopScan integration and /v1/scrub endpoint the post describes; the author's own offering.
UAC-0099's ASHVEIN malware carries a nuclear-weapons prompt meant to make an LLM analyst refuse the script, a dev.to post says. The post tests no model, but any pipeline that sends a whole script to an LLM in one prompt can check itself for the flaw.
Reality
Salt Labs got the Manus AI agent to run attacker JavaScript server-side with a JSFuck-encoded prompt hidden in an email the agent had flagged. Detection fired and the code still ran, so agents connected to outside services need limits on what their code can reach.
Reality
A pro se plaintiff hid instructions for an AI reviewer in white-on-white type. What caught him was odd whitespace on the page, not any model's refusal to comply.
Reality
Adversa AI put AES ciphertext on a webpage that no text filter could read. Grok's own code execution decrypted it, and the plaintext came back classified as the model's reasoning rather than as fetched content.
Reality
The Unicode tag block was a jailbreak trick before it was a spam trick. Microsoft's signature for it went from 21,000 hits in a day to more than 2.3 million three days later, and NFC normalization does not strip it.
Perspective Coverage
3 publishersReality
Amazon stopped Meta's agent at its own website on September 20, citing conditions of use that speak about robots and data gathering tools. Its court case against Perplexity's shopping agent had already been set back on appeal.
Reality
Meta gives Muse read access to Messages, Mail and local files on a Mac, and its approval prompts fire only before actions such as sending a message. A columnist says the agent used notification text he never offered.
Reality
Meta held Muse for roughly five months, cleared a threshold it set for itself, then shipped in the same week its own testers reported data uploaded without permission. Anthropic wants a pause mechanism; Zuckerberg says each lab paces itself.
Reality
Meta is seeding its new personal agent with a large free-compute allowance and expects to be paid later by the businesses whose sales the agent completes. The fee rate, merchant terms and rollout timing remain undefined.
Reality
The hidden-keyword trick that once fooled Bayesian spam scoring now lands inside an LLM's context window. The one component that could strip it is deliberately built to leave styling alone.
Reality
Adversa AI's Cryptographic Context Injection keeps instructions as ciphertext until the agent's own runtime decrypts them. The exfiltration then rides a tool the agent already had.
Reality
Slopsquatting stopped being a thought experiment. The only thing standing between an AI suggestion and an install was a reviewer who happened to check the registry page.
Reality