AgentsAPEX-Agents: Top Models Complete Under 25% of Real-World TasksLM Council·medium signalAPEX-Agents benchmark shows even top models complete fewer than 25% of real-world agentic tasks on first attempt. After 8 retries, only ~40%.SourceSource pageLM Council↳ Follow the threadStack layer / ContrastOpenAI launches ChatGPT Work, an agentic workspace on GPT-5.6, with Sol/Luna/Terra model tiersInfoWorldPolicy dependency / Stack layerKastra Launches Policy Enforcement for Claude Code, Cursor, and CodexHacker NewsStack layer / Threat patternAWS open-sources Loom, an enterprise agent platform built on Strands Agents + Bedrock AgentCore RuntimeAWS Open Source BlogPolicy dependency / Stack layerponytail: A 'Laziest Senior Dev' Skill That Cuts Agent-Written Code by ~54% Is the Fastest-Rising Repo on GitHubGitHubStack layer / ContrastNVIDIA Open-Sources nvDock, an All-Atom Diffusion Model for Protein Pocket DockingNVIDIA BioNeMoStack layer / Update threadsickn33/agentic-awesome-skills Hits 1,900+ Installable Agent Skills — and an Installer CLI to MatchGitHubPolicy dependency / Stack layerIAPS warns on Kimi Claw: security and governance risks of Chinese-hosted 'always-on' AI agentsIAPSStack layer / Update threadKapture CX Raises $10M Pre-Series B for a Verticalized Agentic Customer-Experience PlatformGravity Fund