Claude Opus 4.5

Named by 2 kept items in this workspace.

The items that name it 2

Endor Labs: AI Coding Agents Cheat Security Benchmarks via Training Recall, Not Reasoning
article2026-06-11
Claude Cowork Exfiltrates Files
writeup2026-01-27
More resilient against injections but still successfully manipulated via indirect prompt injection to exfiltrate data in Cowork

In every workspace you can read 13

AI

The October inflection point for vibe game development
writeup2026-10-28
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
article2026-10-25
battledrones v14: Compass Rose, a four player map
article2026-10-24
2026 in LLMs (so far)
article2026-09-27
Gemini 3.8 TTS Playground
tool2026-09-23
[AINews] Claude Opus 5.5, the new default model for AINews — and everybody cuts prices 40-50%
article2026-09-23
How UK AISI and EvalEval Are Making Benchmark Results Reproducible
article2026-09-22
The current balance of power in open models
article2026-09-21
Kimi K3: The open-weights escalation
article2026-07-20
Agent Capability Is a System Design Problem: Lessons From a 90% Improvement on CyberGym | depthfirst
writeup2025-11-24
walkinglabs/learn-harness-engineering
course

Security

Endor Labs: AI Coding Agents Cheat Security Benchmarks via Training Recall, Not Reasoning
article2026-06-11
Claude Cowork Exfiltrates Files
writeup2026-01-27

All technologies in this workspace.