Continue

Mentions · 7d0
Mentions · 30d1
All mentions3
First seenSep 9, 2026
Last mentioned9d ago

Timeline

3 articles · newest first

Context compaction in coding agents can launder forged user instructions

Flattening harnesses accept forged approvals 87-100% of the time, while four role-typed harnesses reject the same attack.

arXiv cs.CROct 21, 15:24 UTC · 9d agoAI safety & security1 · 1 read

Hackers Target Claude, Cursor and Codex AI Agents to Steal Tokens and Prompt Histories

Gen Digital found infostealers like Amatera and Remus stealing AI coding agent tokens, prompt histories, and MCP configs from infected Windows and macOS machines.

Cyber Security NewsSep 9, 11:41 UTC · Sep 9, 2026MalwareExploited5

Infostealers Target Claude, Cursor, Codex and Other AI Agents to Steal Credentials and Sensitive Data

Gen Digital researchers report infostealer families Amatera, Remus and CallbackBeaver now harvest Claude, Cursor and Codex agent data, including tokens and MCP configs.

GBHackersSep 9, 07:05 UTC · Sep 9, 2026MalwareExploited4 · 1 read

Appears with

opencodeproduct · 3
Codexproduct · 3
Clineproduct · 3
Remusmalware · 2
Gen Digitalvendor · 2
Otter Stealermalware · 2
BeeStealermalware · 2
Djinn Stealermalware · 2
APEX Stealermalware · 2
Amateramalware · 2
Cursorproduct · 2
STG Stealermalware · 2
CallbackBeavermalware · 2
Claudeproduct · 2
HydraStealermalware · 2
OpenHandsproduct · 1

Entities are extracted by the model from each article. First seen Sep 9, 2026, last mentioned 9d ago. Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.