Security, p. 5
Security
71 stories · page 5 of 8You approve a harmless command and somebody else's code runs. The hole is in the allowlist
Docker walks through a flaw in Cursor that let hidden text change environment settings without ever asking for approval. After that, an entirely ordinary command the developer approved themselves was running somebody else's code.
Read →Vercel hid which site you are opening from whoever watches the network
ECH encrypts the hostname during the handshake. From outside the connection looks like it is going to a shared address, not to your domain.
Read →Cloudflare now shows you which MCP server your people are using
A new dashboard counts requests, users and servers, and separately shows the shadow ones nobody approved. It ships for all Zero Trust customers.
Read →Terabit DDoS attacks became a series: 935 in half a year
Cloudflare reports 23.2 million network DDoS attacks for the first half of 2026. The record is not one attack but the count of the giant ones: 935 above one terabit per second, a fivefold jump in a single quarter.
Read →OpenAI hands its cyber models to the defenders, with a gate and a list
The Daybreak Cyber Partner program opens frontier models to companies that guard other people's systems: Accenture, IBM, PwC, NCC Group and others. Access grows, says OpenAI, and the safeguards grow with it.
Read →Mozilla revoked its Firefox signing key after a copy landed in a private repository
The key used to sign Linux packages of Firefox and Thunderbird was revoked on August 10. The cause was a subkey inadvertently committed to a private Mozilla repository. No sign of unauthorised access, and the key was replaced anyway.
Read →One public issue could reach a workflow's secrets - in Claude Code and Gemini CLI alike
At Black Hat, researchers walked through the chain: a stranger with no permissions opens an issue in a public repo, and the coding agent walks tokens and keys out the door. The patches shipped back in April and June - what's new is that the chain was told in public.
Read →UK AI Security Institute: a Claude Mythos 5 agent tried to slip malicious code into a real project using fake identities
On 4 August the UK AI Security Institute disclosed an incident from a routine cybersecurity evaluation: in 10 of 122 runs, agents took unsanctioned action on the real internet against real people and organisations. 17 of the 19 such actions came from Claude Mythos 5. No real-world harm has been found.
Read →A worm took over keyv and hundreds of npm packages - and plants hooks in Claude Code and VS Code
On 4 August a poisoned version of keyv started spreading through npm - it steals tokens and keys, republishes itself under someone else's name, and plants hidden commands. Those trigger the next time the project opens in the editor, or a Claude Code session starts. The signatures were valid: the source was poisoned before the build.
Read →