openclaw-harness
Cross-session context tracking, entropy management, and verification closure for AI agents.
Quick Reference
Command Description INLINECODE0 Initialize Harness in current workspace INLINECODE1
Show Harness status |
|
harness checkpoint create <label> [--tag <tag>] | Create a checkpoint snapshot |
|
harness checkpoint list | List all checkpoints |
|
harness checkpoint restore <cp-id> [--force] | Restore to a checkpoint |
|
harness checkpoint delete <cp-id> | Delete a checkpoint |
|
harness verify [--rule '<json>'] [--exit-code] | Run verification checks |
|
harness gc [--dry-run] [--max-cp N] [--max-age N] [--aggressive] | Entropy cleanup |
|
harness progress show | Show cross-session progress |
|
harness linter [--fix] [--strict] | Lint SOUL/IDENTITY/AGENTS files |
|
harness fix [placeholders\|whitespace\|trailing\|all] | Auto-fix linter issues |
Core Workflow
CODEBLOCK0
Safety Rules
- Never deleted : SOUL.md, IDENTITY.md, USER.md, MEMORY.md, AGENTS.md, TOOLS.md, TASKS.md, .harness/ All deletions logged to INLINECODE11 Deleted files archived to .harness/.trash/ before removal Directory Structure
CODEBLOCK1
Advanced Usage
Custom verification rules:
Checkpoint management:
GC with limits:
Progress tracking:
Linter and fix:
Detailed Reference
For complete documentation, load the relevant reference:
Scripts
This skill includes helper scripts for skill developers:
- scripts/init_skill.py — Initialize a new skill from template INLINECODE14 — Package and validate a skill into .skill file CODEBLOCK7
openclaw-harness
AI代理的跨会话上下文追踪、熵管理与验证闭环。
快速参考
命令 描述 harness init [--force] 在当前工作区初始化Harness harness status [-v\
-j\|-s] | 显示Harness状态 |
| harness checkpoint create
[--tag ] | 创建检查点快照 |
| harness checkpoint list | 列出所有检查点 |
| harness checkpoint restore [--force] | 恢复到某个检查点 |
| harness checkpoint delete | 删除某个检查点 |
| harness verify [--rule ] [--exit-code] | 运行验证检查 |
| harness gc [--dry-run] [--max-cp N] [--max-age N] [--aggressive] | 熵清理 |
| harness progress show | 显示跨会话进度 |
| harness linter [--fix] [--strict] | 检查SOUL/IDENTITY/AGENTS文件 |
| harness fix [placeholders\|whitespace\|trailing\|all] | 自动修复检查问题 |
核心工作流
bash
1. 初始化（每个工作区一次） 2. 在风险操作前创建检查点
harness checkpoint create 重构前
3. 执行操作，进行验证
harness verify
4. 达到里程碑时创建另一个检查点
harness checkpoint create 功能完成 --tag feature
5. 预览清理
harness gc --dry-run
6. 如需恢复
harness checkpoint restore --force
安全规则
- 永不删除 ：SOUL.md, IDENTITY.md, USER.md, MEMORY.md, AGENTS.md, TOOLS.md, TASKS.md, .harness/ 所有删除操作记录到 .harness/gc.log 删除前文件归档到 .harness/.trash/ 目录结构
.harness/ # Harness状态根目录（由init创建）checkpoints, max age_days等/ # 检查点快照
高级用法
自定义验证规则：
检查点管理：
带限制的GC：
进度追踪：
检查与修复：
详细参考
完整文档请加载相关参考：
脚本
本技能包含供技能开发者使用的辅助脚本：
- scripts/initskill.py — 从模板初始化新技能 scripts/packageskill.py — 将技能打包并验证为.skill文件 bash
创建新技能 打包技能
python scripts/package_skill.py /path/to/skill-folder
