self-evolving-agent
An OpenClaw skill that upgrades self-improving agents from reactive error logging to goal-driven capability evolution with curriculum, evaluation, transfer, and promotion.
github
skill
agent
ai-agent
openclaw
self-improving-agent
python
⭐ 8⬇️ 0🏷️ main📦 2
self-evolving-agent-main.zip
84.2 KB下载
安装与使用说明
Codex
- 先下载 ZIP 包（例如：self-evolving-agent-main.zip）。
- 解压后放到 $CODEX_HOME/skills/ 目录下。
- 如果没有设置 $CODEX_HOME，通常可放到 ~/.codex/skills/。
- 确保目录内有 SKILL.md，然后重开会话即可使用。
源码目录预览
├── .github │ ├── ISSUE_TEMPLATE │ │ ├── bug_report.yml │ │ ├── config.yml │ │ └── feature_request.yml │ ├── workflows │ │ └── ci.yml │ └── pull_request_template.md ├── agents │ └── openai.yaml ├── assets │ ├── index │ │ └── manifest.json │ ├── records │ │ ├── agenda │ │ │ └── agd-bootstrap-001-bootstrap-agenda.md │ │ └── capabilities │ │ ├── cap-bootstrap-001-research.md │ │ ├── cap-bootstrap-002-planning.md │ │ ├── cap-bootstrap-003-tool-use.md │ │ ├── cap-bootstrap-004-verification.md │ │ ├── cap-bootstrap-005-synthesis.md │ │ ├── cap-bootstrap-006-communication.md │ │ ├── cap-bootstrap-007-coding.md │ │ ├── cap-bootstrap-008-execution-discipline.md │ │ ├── cap-bootstrap-009-memory-retrieval.md │ │ └── cap-bootstrap-010-long-horizon-task-handling.md │ ├── CAPABILITIES.md │ ├── ERRORS.md │ ├── EVALUATIONS.md │ ├── FEATURE_REQUESTS.md │ ├── LEARNINGS.md │ ├── LEARNING_AGENDA.md │ └── TRAINING_UNITS.md ├── benchmarks │ ├── schemas │ │ └── judge-output.schema.json │ └── suite.json ├── demos │ ├── demo-1-diagnosis.md │ ├── demo-2-training-loop.md │ ├── demo-3-promotion-and-transfer.md │ ├── demo-4-agenda-review.md │ └── demo-5-pre-task-risk-diagnosis.md ├── evals │ └── evals.json ├── hooks │ └── openclaw │ ├── HOOK.md │ └── handler.ts ├── modules │ ├── capability-map.md │ ├── curriculum.md │ ├── diagnose.md │ ├── evaluator.md │ ├── learning-agenda.md │ ├── promotion.md │ └── reflection.md ├── scripts │ ├── activator.sh │ ├── bootstrap-workspace.sh │ ├── error-detector.sh │ ├── evolution_runtime.py │ ├── migrate-self-improving.py │ ├── run-benchmark.py │ └── run-evals.py ├── system │ └── coordinator.md ├── .gitignore ├── CHANGELOG.md ├── CONTRIBUTING.md ├── LICENSE ├── README.md ├── README.zh-CN.md ├── SECURITY.md ├── SKILL.md └── install.md
元数据
Slug: self-evolving-agent
作者: RangeKing
协议: MIT
包大小: 84.2 KB
SHA256: d7b6226671eb231be5cc3d12460b68346124dab6e09bb95c322fb82125a3b650