Anthropic released Strands harness, a general-purpose agent framework available under Apache 2.0 that runs locally or in the cloud. It achieves 28% lower costs than comparable Claude and GPT harnesses while maintaining equal or better accuracy through optimized prompt caching and context management. The framework supports multiple model providers and includes built-in tools for shell, file, and web operations.
Unreal Labs presents an async-first agent harness library with components for managing LLM agent sessions, tool calls, and asynchronous operations. The system uses event-driven architecture with session-scoped deduplication, append-only persisted history, and composable tool translators that validate and convert model requests into serializable operations.
A developer tested five coding agents on the same local model (qwen3-coder-next) and frozen test suite, finding that 90% of failures stemmed from harness problems rather than model limitations. The model generated correct code ~97-98% of the time, but agents failed to properly execute, verify, or complete tasks due to tooling issues like hard-coded turn limits and poor stopping conditions.
RRSI is a method for automatically improving LLM agent systems by iteratively refining prompts, control flow, and tooling while preventing overfitting through regularization constraints. The approach uses a budget-limited proposer and a critic-pruner selector to favor reusable mechanisms, achieving significant gains on in-distribution and out-of-distribution benchmarks while reducing computational cost.
Apprentice is a highly configurable Common Lisp coding harness for AI agents, built around abstractions for models, tools, anchors, and loops. It supports multiple LLM providers and enables rapid customization through Lisp macros and REPL-driven development.
Cayu is a Python framework for building domain-specific AI agent harnesses. The article demonstrates assembling a campus dining purchase agent by defining the job, creating an agent identity, building tools that wrap business logic, and providing policies that guide the model's decision-making process.