Fetching from the wire…
01
02
03
04
05
06
07
08
Agents2026-08-09 · source-backed
AMAP-ML released it August 4 (463 stars, v0.1.3 on August 7) on a strict Manager/Executor/Auditor split where, in its own words, "only results that pass independent verification enter persistent task state" (GitHub). WeaveBench 51.8% → 80.7% completion. OSWorld 2.0 2.8% → 8.3% full completion. Terminal-Bench 2.1 69.7% → 77.2% using 24% fewer tokens. It wraps Codex and Claude Code via an AgentAdapter protocol, defaults to 30 rounds, and persists goals, verified progress, audit reports, and role trajectories in isolated run directories. Same architectural lesson as HTTP Terminator: the auditor role, not a bigger context window, is what stops multi-hour runs from drifting.
Each link below shares sources, entities, or timing with this story.
AMAP-ML/LongHorizon-Harness (370 stars, MIT, paper at arXiv 2608.01964) splits long computer-use work into three roles with independently assignable models. Reported: WeaveBench 51.8% → 80.7%, Terminal-Bench 2.1 69.7% → 77.2%, OSWorld 2.0 2.8% → 8.3%. It integrates with Claude...
A Chinese lab shipped a runtime that manages two American coding agents as subagents, and it went from repo creation to 145,439 stars in four days. deepseek-ai/deepseek-harness published dsh-v0.1.0-rc.7 at 12:01 UTC today, its first tagged release since the repo appeared on Au...
Three moves, two days, no coordination between them. August 10–11: GitHub shipped Ollama as a BYOK provider inside Copilot for JetBrains (GitHub Changelog). Unsloth released Unsloth Desktop with a command literally named unsloth start claude, which points Claude Code and Codex...
A spec is a press release until someone who didn't write it implements it. GitHub made Agent Plugins 1.0 generally available on August 12 across VS Code, Copilot CLI, the Copilot SDK, and the Copilot app on all plans. The spec, published August 6, was co-authored by AWS, Anysp...
Three things happened this month that only make sense together. Agent Plugins 1.0 shipped co-signed by six competitors: AWS, Anysphere, Microsoft, OpenAI, Vercel and Google (GitHub Changelog). It makes skills-plus-MCP bundles portable across clients. OpenAI's August 11 Codex c...
312 stars and 24 forks within 24 hours of appearing August 11, MIT (GitHub). Notable for targeting three harnesses at once, which puts it in the emerging agent-agnostic UI layer category rather than single-vendor plugin territory. One day old, no releases, no independent verif...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.