agentsclimarketplace

Debugging

Skill DearCaat/research-foundation-skills/skills/debugging

系统调试:遇到 bug、报错、测试失败、性能退化、意外实验结果或异常数据时,先于任何修复动作使用;仅背景不清而无故障信号时用 grounding,正常执行既定计划时用 executing-plans。From its SKILL.md

Install
npx -y skills add DearCaat/research-foundation-skills --skill debugging

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.2 KB, as published. Nobody here has run it

debugging 系统调试

目标:以 tight 反馈回路查明根因;systematic is faster than thrashing。

Iron Law

根因有证据前不得提出修复;只处理症状即失败。时间越紧,越要系统查证而非猜测。

做法

  • 先建立 tight 反馈回路:找到一条能让此问题变红的快速、确定、可重复检验命令;回路建成,问题已解决 90%。
  • 非确定性问题不强求干净复现,先把复现率提高到足以调试。
  • 意外实验结果同样适用:先复现、排除管线错误、单变量归因,再讨论科学解释。

Gate

复现 → 最小化 → 单变量假设逐一验证 → 修复 → 回归验证。先证明检验能对此问题变红,再证明修复后变绿;gate 未过不得前进。

无法建立回路时,明说并列出已尝试项,向用户索要所需环境、数据或许可;问而不猜。

完成判据

根因有证据支撑;修复后回路变绿;回归检验已留存。

Keep looking

Skills are one crate of 610,045. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.