Partially Verified65% confidenceFactExact time
The core idea of Harness2 is to recursively rewrite the agent harness at test time without updating the model weights.
3
Sources
65%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedHarness2's working loop is: the Agent proposes modifications to the existing Harness for a new task, runs several candidate versions in parallel, and selects the better one by comparing actual performance.79% similarVerifiedThe concept of Harness originates from Test Harness in software engineering, and in the AI Agent context it is extended to mean the systematic engineering that enables an Agent to run continuously, cyclically, and in a controlled manner78% similarUnverifiedAgent harnesses differ from prompt engineering: the latter only affects a single model input, whereas the harness determines how an agent keeps operating across multiple turns and multiple tools.76% similarUnverifiedThe paper "Self-Evolving Agent Harness" proposes achieving self-evolution by revising the agent's software harness while keeping the model weights unchanged.73% similarUnverifiedThe Agent Harness does not replace the reasoning capability of the core model, but instead builds a runtime management shell around the model, responsible for cross-cutting concerns such as tool invocation, state management, and error recovery.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1060116API
curl https://kongchang.com/api/v1/knowledge/claims/1060116MCP
get_claim(id=1060116)