Agent harnessAI and agent development

self-evolve

Iterate agent skills with deterministic acceptance and rollback gates

Stars
0
Forks
0
License
MIT
Updated
Updated Jun 26, 2026

Checking live repository facts…

Project overview

self-evolve provides a methodology and deterministic harness for improving a skill, repository, or project over repeated cycles. The loop reflects, proposes, evaluates, judges, and accepts changes, while code—not the model—decides acceptance, rejection, rollback, and which signal source to trust. It supports program checks, anchor verification, and generative evaluation for open-ended work where ground truth is limited.

Repository facts

Primary language
Python
License
MIT
Repository updated
Jun 26, 2026
Default branch
main

Resource types

Agent harnessGeneral skillPlugin

Use cases

AI and agent development

Platforms

Claude CodeCodex

Runtime

Command lineLocalSandboxed

Capabilities

Security guardrailVerification and evals

Related projects

Browse more similar projects