self-modification

dsh 自托管运维守护：agent 改完代码先验证构建测试再重启，起不来自动回滚。镜像仓，与 Khorsheed/dsh-plugins 的 packages/ankh-guard 自动同步。

An advanced Rust-native AI coding assistant that provides semantic search, impact analysis, automated planning, and safe self-modification capabilities including context packaging, documentation generation, memory management, and Git-integrated workflows.

Experimental autonomous AI agent — memory architecture, IIT Phi consciousness, recursive self-improvement, multi-agent systems.

AI safety theory of infinite deferral: postponing self-modification, pursuing resources and recursive self-improvement, and risks of betrayal and resistance to human control.

.net driven AI self-modification tool with version management. These modules are capable of testing, rolling back, and disseminating seamless updates as they become available through targeted self-learning.

Self-referential Darwin-style agent forge that evolves its own mutation operators, evaluators, and core components. Built with DeepSeek V4.

Grounded self-diagnosis and safe self-modification for OpenClaw agents — no hallucinated field names, no wrong-agent writes, no stale docs

Replay-certified self-modification PoC: public-evidence admission gate with generated comparator covers, lower-tail checks, safe/unsafe growth fallback, and target-witness diagnostics.

therain2020-agent — self-healing, self-evolving AI agent. Modifies its own source code to fix bugs and autonomously grows new capabilities (skills, tools, engine). TypeScript + Bun + React/Ink.

Ethics circuit breaker for self-modifying agents: admission control and alignment-cost measurement | 自进化智能体的伦理熔断机制：面向自我修改的准入控制与对齐成本测量（研究原型）