Execution Harness — 导航入口#

40 patterns x 6 轴，让 Claude Code agent 把活干完。不是框架，不做模型调用——只管 hook 和脚本。

这个 skill 不执行操作。它帮你找到该用哪个子 skill。

你遇到什么问题？#

问题用哪个 skill关键 patternAgent 做了一半就停了execution-loop1.1 Ralph Stop Hook — 阻止提前退出，5 个安全阀"应该可以"但没跑测试execution-loop1.2 Doubt Gate — 检测投机语言，要求验证cargo build 重试 12 次tool-governance2.1 Tool Error Escalation — 3 次提示、5 次 blockrm -rf 毁了未提交代码tool-governance2.3 Checkpoint + Rollback — 自动 git stash压缩后忘了设计决策context-memory3.1 Handoff Documents — 决策写磁盘Context 快满了还在读大文件context-memory3.4 Token Budget — 80%+ 禁止直读限速后 tmux 挂死error-recovery5.1 Rate Limit Recovery — cron 扫描恢复5 个 agent 编辑同一文件multi-agent4.3 File Claim and Lock — 10min TTL 排他锁提交了编译不过的代码quality-verification6.4 Test-Before-Commit — commit 前跑测试

6 轴速查#

轴SkillPattern 数核心能力1execution-loop7Ralph Stop Hook、Doubt Gate、Drift Re-anchoring2tool-governance6错误升级、权限否决、破坏性命令拦截3context-memory8Handoff 文档、Compaction 抢救、Token Budget4multi-agent6Coordinator/Fork/Swarm、文件锁、盲审分离5error-recovery7限速恢复、Crash 恢复、模型降级建议6quality-verification6编辑后 lint、commit 前测试、session 指标

最小配置#

3 个 hook 解决最常见的 3 个问题：

jsonc

// ~/.claude/settings.json
{
  "hooks": {
    "Stop": [{"hooks": [
      {"type": "command", "command": "bash execution-loop/scripts/ralph-stop-hook.sh"},
      {"type": "command", "command": "bash execution-loop/scripts/doubt-gate.sh"}
    ]}],
    "PreToolUse": [{"hooks": [
      {"type": "command", "command": "bash tool-governance/scripts/tool-error-advisor.sh"}
    ]}],
    "PostToolUse": [{"matcher": {"tool_name": "Write|Edit|MultiEdit"}, "hooks": [
      {"type": "command", "command": "bash quality-verification/scripts/post-edit-check.sh"}
    ]}]
  }
}

仓库#

GitHub: lanyasheng/execution-harness | 90 tests | 依赖：bash、jq、python3、pytest