BuildAutomaton open-sources meta-harness agent framework

// 1h agoOPENSOURCE RELEASE

BuildAutomaton open-sources meta-harness agent framework

BuildAutomaton announced the open-source release of meta-harness, a monorepo framework built on the Agent Client Protocol (ACP) to orchestrate and switch between multiple AI agent harnesses across arbitrary compute infrastructure. The repository is rolling out incrementally starting with the @buildautomaton/agent-runtime package, which enables local agent discovery, capability probing, prompt execution, and real-time streaming of plans, todos, permissions, and file changes.

// ANALYSIS

As the landscape of AI agents fragments across competing frameworks, meta-harness smartly tackles runtime interoperability by building on open protocols and CLI bridging rather than creating yet another walled garden. It enables developers to dynamically swap agent harnesses based on capability, cost, and task requirements instead of locking into a single proprietary stack. Implementing the Agent Client Protocol standardizes how agents stream plans, manage permissions, and report file mutations, while the CLI bridge architecture allows deployment across self-hosted or local compute rather than depending on proprietary cloud backends. However, because the codebase is rolling out incrementally in a pre-official release phase, broader ecosystem adoption and stability will take time to materialize.

// TAGS

agentagent-harnessagent-client-protocolopen-sourcedevtooltypescriptruntime

DISCOVERED

1h ago

2026-09-11

PUBLISHED

1d ago

2026-09-09

RELEVANCE

7/ 10

AUTHOR

rchodava

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

TUTORIAL1h ago

Helmryth creator argues against multi-agent bloat

In an architectural deep dive, Helmryth creator Divyam Talwar challenges the assumption that adding more agents linearly scales judgment, arguing that extra responders multiply routing overhead and verification debt. Using examples from Helmryth, he outlines explicit routing contracts—named leads, selected specialists, and full crews—to enforce single-owner accountability over superficial team expansion.

MODEL2h ago

NASA and IBM release Lunar Foundation Model

NASA and IBM have publicly released the NASA-IBM Lunar Foundation Model, an open-source geospatial AI system designed specifically for planetary science and lunar exploration. Developed under the Prithvi model family and made available on Hugging Face and GitHub, the model was trained on SomBench—a comprehensive dataset aggregating data from nine instruments across four missions, notably spanning 17 years of Lunar Reconnaissance Orbiter (LRO) observations across more than 30 spatially aligned data layers. The system outperforms traditional mapping techniques by up to 23% in accuracy, enabling automated identification of water-ice deposits in permanently shadowed regions, geological features like volcanic patches, and crater hazards to assist mission planning for NASA's Artemis program.

NEWS2h ago

OCI revenue surges 121% with $664B backlog

Oracle's AI cloud business experienced rapid growth with Oracle Cloud Infrastructure (OCI) revenue surging 121% year-over-year and its total contract backlog reaching $664 billion. However, the aggressive infrastructure expansion required $28.5 billion in quarterly capital expenditures, pushing free cash flow into negative territory and sparking debate over how quickly these heavy AI investments will translate into sustained shareholder returns.

BuildAutomaton open-sources meta-harness agent framework