MemoHarness adapts agents using execution history
Published on July 17, 2026, the work presents a method that adjusts six dimensions of an agent harness using prior executions. On a shell-agent benchmark, it scored 0.806, compared with 0.722 for the strongest fixed-harness baseline.
Source: MemoHarness adapta harnesses de agentes usando histórico de execuções (arxiv.org). Text prepared with AI from this source.
What happened and what to do
Published on July 17, 2026, the work on MemoHarness describes a method that adapts six dimensions of agent harnesses and retrieves lessons from similar past cases. On a shell-agent benchmark, it scored 0.806, compared with 0.722 for the strongest fixed-harness baseline. The method uses agents’ own executions, without test-time labels or additional search.
A company can assess this approach in shell-agent workflows by logging executions, grouping similar cases, and measuring whether harness adjustments improve representative tasks. An implementation could include execution collection and monitoring, evaluation criteria, and controls to review changes before applying them in production.
How the consultancy can help
Wendelmaques can diagnose the agent workflow and its risks, define an evaluation scope and metrics, and implement execution collection, monitoring, and controlled harness adjustments, along with the agreed operation.
Next step
Send a short description of your agent workflow and desired outcome to receive a scoped proposal.
Let's assess your project
I reply with scope and a proposal.