{"version":1,"type":"story","url":"https://digestai.news/story/google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo","json":"https://digestai.news/story/google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo.json","markdown":"https://digestai.news/story/google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo.md","slug":"google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo","headline":"Google Cloud AI Research open-sources Harness² recursive agent framework","summary":"Google Cloud AI Research released Harness², an open‑source recursive agent harnessing framework version 0.1.0. The framework adds eight editable harness components—system prompt, guardrails, memory, tool guidance, subagents, skills, scripts, and plugins—while keeping model weights frozen. It runs on Linux with Python 3.12 or 3.13 and uses Vertex AI to call models; GPU requirements are not disclosed.\n\nThe research team evaluated Harness² on three professional benchmarks (WorkBuddy‑Bench, JobBench, Harvey’s LAB) using Gemini 3.5 Flash, GLM‑5 under OpenCode and Codex, with Claude Opus 4.8 as the default editor. A single refinement step lifted GLM‑5 (OpenCode) from 64.2 to 75.7 on WorkBuddy‑Bench (+11.5 points), while the same model under Codex improved only 1.2 points (23.6 to 24.8). Parallel recursion with k=5 raised Gemini (OpenCode) to 80.8 on WorkBuddy‑Bench. The framework supports both parallel and sequential recursion modes, storing edits in an EXPERIENCE.md library for future tasks.","keyPoints":["Harness² provides eight editable harness components and refines agents without changing model weights.","One refinement step raised GLM‑5 (OpenCode) from 64.2 to 75.7 on WorkBuddy‑Bench (+11.5 points).","Runs on Linux with Python 3.12/3.13, uses Vertex AI for model calls; GPU needs not disclosed."],"whyItMatters":"Enables developers to improve agent performance at deployment without costly model retraining, facilitating rapid adaptation to unseen tasks.","category":{"slug":"agents","name":"Agents & Tools","url":"https://digestai.news/category/agents"},"entities":{"companies":["Google","Google Cloud AI Research"],"models":["Gemini 3.5 Flash","GLM-5","OpenCode","Codex","Claude Opus 4.8"],"people":["Asif Razzaq"]},"firstPublishedAt":"2026-10-18T08:15:04Z","updatedAt":"2026-10-18T08:15:04Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"MarkTechPost","title":"Google Open-Sources Harness2, a Recursive Agent Harnessing Framework","url":"https://marktechpost.com/2026/10/18/google-harness2-recursive-agent-harnessing-framework","publishedAt":"2026-10-18T08:15:04Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"Google's Quest to Tame Self‑Improving AI","url":"https://digestai.news/thread/google-research-open-sources-rrsi-for-self-improving-ai-agents-without","storyCount":3},"cite":{"text":"Digest AI, \"Google Cloud AI Research open-sources Harness² recursive agent framework\", 18 October 2026, https://digestai.news/story/google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo","publisher":"Digest AI","title":"Google Cloud AI Research open-sources Harness² recursive agent framework","datePublished":"2026-10-18T08:15:04Z","url":"https://digestai.news/story/google-cloud-ai-research-open-sources-harness2-recursive-agent-framewo"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}