Skip to content
Agentmart
Submit a tool

systematic-debugging

Use when a specific failure, error, or flaky test needs its root cause proven before fixing—not for applying known fixes.

Skill by shiyuan17

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/shiyuan17-systematic-debugging

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when a specific failure, error, or flaky test needs its root cause proven before fixing—not for applying known fixes.

Use cases: testing

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Shell scripts

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/systematic-debugging

What's inside

Skill name
systematic-debugging
Description
Use when a specific failure, error, or flaky test needs its root cause proven before fixing—not for applying known fixes.
License
MIT
Instructions
20 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (4)

  • SKILL.md
  • openai.yaml · agents/
  • find-polluter.sh
  • metadata.json

SKILL.md (first lines)

# 系统定位故障

先证明根因，再修改代码。

## 执行

1. 稳定复现症状，记录最小输入、环境、期望和实际结果。
2. 定位直接失败点，并沿数据或调用链追踪到最早错误状态。
3. 若存在多个合理原因，列出 2–3 个竞争性假设；每个假设记录支持证据、反证和能区分它与其他假设的最小实验。根因已由证据唯一确定时，不强行制造额外假设。
4. 一次只改变一个主要变量并收集证据。同一症状经过两次修复仍失败，且没有新的区分性证据时，暂停扩大改动，核对已尝试方案与反证，再执行一个能区分剩余假设的最小实验。这是诊断触发器，不是硬门禁或自动终止；根因已明确则继续修复，基础设施重试不计入修复次数，不强制复盘文档或独立 Agent。
5. 用失败测试或最小复现固定问题，实施最小修复。
6. 运行聚焦回归及相关套件，确认测试没有弱化。

怀疑测试顺序污染时运行 `find-polluter.sh <file_or_dir_to_check> <test_pattern> [test_command_template]`：测试模式从当前目录解释（`./` 前缀可选），`**` 匹配零级或多级目录；`{file}` 必须是独立参数且最多出现一次，模板只支持字面量引号分组，不接受 shell 展开、管道、重定向和命令串联（默认模板 `npm test {file}`）。退出码：`0` 表示每个候选都跑过且未产生污染，`1` 表示已定位污染者，`2` 表示没有得出结论（参数非法、模板不支持、零候选、运行前污染已存在或测试命令失败），不能当作 clean 使用。不要用固定 sleep、批量猜改或重复尝试掩盖未知根因。

## 交付

报告复现、候选假设及其排除证据、根因证据、最小修复、回归结果和仍未排除的风险。归因与返修记录沿用 `docs/rules/troubleshooting.md` 的收尾口径，不另建学习系统。

## 正反例

- 该用：「登录偶发 500，日志只留一次堆栈」——根因未定，需要能推翻判断的区分实验。
- 根因已知：证据已唯一确定根因，或只是实施既有结论的机械修复，直接改并跑聚焦回归，不重新制造假设。
- 工具不可用：无法运行复现时只交付静态推导与下一步实验，明确标注「未复现」，不声称已证明根因。

From shiyuan17/Vibe-Harness, MIT.

More from shiyuan17/Vibe-Harness

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Oct 14, 2026 (15 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
Repository
github.com/shiyuan17/Vibe-Harness
First seen
Oct 24, 2026
Last checked
Oct 24, 2026
Source
GitHubgithub-graphql

More from shiyuan17

Cite this page

systematic-debugging skill. Agent Mart. https://agentmart.dev/skills/shiyuan17-systematic-debugging (last updated 2026-10-24)