Security
Anthropic will not change MCP command execution, OX Security finds
OX Security tells that the protocol starts command text with no check, and that Anthropic confirmed this is by design.
forkast.newsClaimed, not confirmed
What matters in AI.
SubscribeTag
240 stories carry this tag.
Security
OX Security tells that the protocol starts command text with no check, and that Anthropic confirmed this is by design.
forkast.newsClaimed, not confirmed
Agents
One change to the harness increased the GLM-5 score by 11.5 points, with no change to the model weights.
marktechpost.comClaimed, not confirmed
Models
At off-peak times, the price for each token can be about 80 percent below the standard rate.
Security
The authors want other groups to work on the 51 open problems in the report.
Agents
In tests, the performance of 4 models increases by 2.9 to 3.7 percentage points.
Agents
The twin tracks well volumes, and the simulator rejects commands that cannot work.
Security
The seller says the tool found 1,072 issues on 199 hosts, but Bitsight cannot check this number.
Agents
The user must tag Chrome in the ChatGPT web or desktop app first.
Research
Belfield, an engineer at Google, reports that AI writes more than 75 percent of production code, but a human examines each code change.
Agents
Google writes that MCP traffic stays in the cluster and that no connection starts from Google Cloud.
Models
Each user gets 2 free document checks each day, and a user with a Gemini API key gets more.
Security
The Chain of Trust gives a check of an agent at each step, from discovery to payment to audit.