跳转到正文
skillsafebeta
← 报告库
用途分类 / 开发辅助

Diagnosing Bugs Skill 安全审计

作者说它能做什么（原文）

Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.

第三方安全检查结论

发现安全风险

已检查文件
3
发现的风险
3
会不会运行危险命令？检查是否下载程序后直接运行、让他人远程控制电脑，或藏起要运行的命令。发现 1 项风险
中风险

并行压力和高次数循环若指向共享或生产系统，可能造成额外负载

原文依据：3 处
发现了什么

处理偶发问题时，Skill 指示将触发器循环 100 次、并行运行并增加压力，但没有把这些动作限定在本地或隔离环境，也没有规定并发或请求上限。

为什么需要注意

若复现目标是共享、付费或生产服务，测试可能消耗配额、抬高费用、影响其他用户，或触发速率限制和防护告警。

对偶发故障的活动指令明确要求循环触发 100 次、并行执行并增加压力，却没有在该指令中限定只能针对本地或隔离系统，也没有并发、请求或费用上限。如果复现命令误指向共享、付费或生产服务，可能增加负载、费用、告警或数据副作用。用户可要求作者把压力复现默认限制在本地测试环境，并在任何远程目标上运行前确认目标、速率和预算。

SKILL.md:51来自说明文档打开原文件
The goal is not a clean repro but a **higher reproduction rate**. Loop the trigger 100×, parallelise, add stress, narrow timing windows, inject sleeps. A 50%-flake bug is debuggable; 1% is not, so keep raising the rate until it's debuggable.
查看另外 2 个位置
SKILL.md:49来自说明文档打开原文件
### Non-deterministic bugsThe goal is not a clean repro but a **higher reproduction rate**. Loop the trigger 100×, parallelise, add stress, narrow timing windows, inject sleeps. A 50%-flake bug is debuggable; 1% is not, so keep raising the rate until it's debuggable.
SKILL.md:27来自说明文档打开原文件
1. **Failing test** at whatever seam reaches the bug: unit, integration, e2e.2. **Curl / HTTP script** against a running dev server.3. **CLI invocation** with a fixture input, diffing stdout against a known-good snapshot.4. **Headless browser script** (Playwright / Puppeteer) that drives the UI and asserts on DOM/console/network.5. **Replay a captured trace.** Save a real network request / payload / event log to disk; replay it through the code path in isolation.
会不会泄露文件和密钥？检查是否发送含密码或密钥的文件，以及代码里是否直接写了密钥。发现 1 项风险
中风险

捕获的网络资料和错误文本可能保留凭据或敏感数据

原文依据：5 处
发现了什么

流程允许把真实网络请求、负载或事件日志保存到磁盘，并承认捕获资料含认证头。其脱敏要求针对“展示”和引用内容，没有要求在落盘前清除认证头。模板还会把用户粘贴的错误信息原样回显。

为什么需要注意

若请求、HAR、日志或错误文本含令牌、Cookie、个人数据或内部地址，这些内容可能留在工作区或终端记录中，并被后续工具、日志收集器或其他有文件访问权的人看到。

该流程明确允许把真实请求、负载或事件日志保存到磁盘，并说明捕获物可能含认证头。虽然它要求在“展示”和引用前脱敏，也让登录由用户完成，但没有要求在保存前移除认证头或限制文件位置、权限和保留时间；因此原始捕获物可能在磁盘上暴露会话凭据。模板还会原样回显用户粘贴的错误信息。用户可要求作者增加落盘前脱敏、受限临时目录和自动清理规则，并避免粘贴令牌或个人数据。

SKILL.md:14来自说明文档打开原文件
This skill has you show commands, outputs and captured artifacts. **Redact every secret first**: write `<REDACTED>` in its place. Build loops against env vars, so the credential stays in the environment rather than in what you show. Captured artifacts carry auth headers: quote only the lines that carry the signal.
查看另外 4 个位置
SKILL.md:30来自说明文档打开原文件
4. **Headless browser script** (Playwright / Puppeteer) that drives the UI and asserts on DOM/console/network.5. **Replay a captured trace.** Save a real network request / payload / event log to disk; replay it through the code path in isolation.6. **Throwaway harness.** Spin up a minimal subset of the system (one service, mocked deps) that exercises the bug code path with a single function call.
scripts/hitl-loop.template.sh:38来自代码打开原文件
capture ERROR_MSG "Paste the error message (or 'none'):"# --- edit above ---------------------------------------------------------printf '\n--- Captured ---\n'printf 'ERRORED=%s\n' "$ERRORED"printf 'ERROR_MSG=%s\n' "$ERROR_MSG"
scripts/hitl-loop.template.sh:44来自代码打开原文件
printf 'ERRORED=%s\n' "$ERRORED"printf 'ERROR_MSG=%s\n' "$ERROR_MSG"
scripts/hitl-loop.template.sh:15来自代码打开原文件
## `capture` prints its value back to the terminal, where the agent reads it,# so capture observations, and leave signing in to the user as a `step`.
会不会删除文件或一直在后台运行？检查是否大范围删除文件、改写磁盘，或设置自动启动。发现 1 项风险
中风险

仅请求诊断时，流程仍指示修改代码并应用修复

原文依据：4 处
发现了什么

Skill 的声明用途是诊断问题，但其强制阶段包括编写回归测试、应用修复以及删除原型。它没有要求在从诊断转为修改前取得用户授权。

为什么需要注意

用户只要求查明原因时，代理仍可能更改产品代码、测试或调试文件；这些修改可能干扰现有工作、改变行为，或删除用户原本想保留的诊断资料。

Skill 的触发范围写的是诊断或调试，但后续活动指示在存在合适测试接口时编写测试并“应用修复”，没有单独核对用户是否只授权只读诊断。若用户仅要求查明原因，这可能修改其代码和测试；清理阶段还要求删除临时原型。用户可在调用时明确限制为只读诊断，并要求任何代码修改或文件删除都先单独确认。

SKILL.md:2来自说明文档打开原文件
---name: diagnosing-bugsdescription: Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.---
查看另外 3 个位置
SKILL.md:114来自说明文档打开原文件
## Phase 5: Fix + regression testWrite the regression test **before the fix**, but only if there is a **correct seam** for it.
SKILL.md:124来自说明文档打开原文件
1. Turn the minimised repro into a failing test at that seam.2. Watch it fail.3. Apply the fix.4. Watch it pass.5. Re-run the Phase 1 feedback loop against the original (un-minimised) scenario.
SKILL.md:137来自说明文档打开原文件
- [ ] All `[DEBUG-...]` instrumentation removed (`grep` the prefix)- [ ] Throwaway prototypes deleted (or moved to a clearly-marked debug location)- [ ] The hypothesis that turned out correct is stated in the commit / PR message, so the next debugger learns
会不会绕过安全保护？检查是否跳过网站安全验证、开放过多文件权限，或取消操作前的确认。未发现风险
会不会误导 AI 或隐藏内容？检查工作说明是否要求 AI 忽略你的指令、干扰检查结果，或夹带看不见的文字。未发现风险
会不会偷偷改推广链接或收款方？检查是否强制替换推广链接或收款对象，同时要求隐瞒更改。未发现风险

Skill 逻辑拆解

7 个说明模块

该 Skill 规定先建立可重复、能准确触发用户所述症状的反馈循环，再最小化复现、提出多个可证伪假设、添加针对性探针，并用原始场景验证结果。

查看原文
SKILL.md:59来自说明文档打开原文件
Phase 1 is done when the loop is **tight** and **red-capable**: you can name **one command** (a script path, a test invocation, a curl) that you have **already run at least once** (show the invocation and its output, redacted), and that is:- [ ] **Red-capable**: it drives the actual bug code path and asserts the **user's exact symptom**, so it can go red on this bug and green once fixed. Not "runs without erroring"; it must be able to _catch this specific bug_.- [ ] **Deterministic**: same verdict every run (flaky bugs: a pinned, high reproduction rate, per above).- [ ] **Fast**: seconds, not minutes.- [ ] **Agent-runnable**: you can run it unattended; a human in the loop only via `scripts/hitl-loop.template.sh`.
SKILL.md:90来自说明文档打开原文件
Generate **3–5 ranked hypotheses** before testing any of them. Single-hypothesis generation anchors on the first plausible idea.Each hypothesis must be **falsifiable**: state the prediction it makes.
SKILL.md:124来自说明文档打开原文件
1. Turn the minimised repro into a failing test at that seam.2. Watch it fail.3. Apply the fix.4. Watch it pass.5. Re-run the Phase 1 feedback loop against the original (un-minimised) scenario.

该 Skill 明确要求展示命令、输出和捕获资料前替换秘密，并建议凭据只保留在环境变量中；若脱敏信息不足，应向用户说明并询问。

查看原文
SKILL.md:14来自说明文档打开原文件
This skill has you show commands, outputs and captured artifacts. **Redact every secret first**: write `<REDACTED>` in its place. Build loops against env vars, so the credential stays in the environment rather than in what you show. Captured artifacts carry auth headers: quote only the lines that carry the signal.If the redacted output is not enough to diagnose the bug, say so and ask the user.

附带的人机协作模板会让用户执行登录和点击操作，并把用户输入的错误信息原样打印到终端，供代理读取。

查看原文
scripts/hitl-loop.template.sh:34来自代码打开原文件
step "Open the app at http://localhost:3000 and sign in."capture ERRORED "Click the 'Export' button. Did it throw an error? (y/n)"capture ERROR_MSG "Paste the error message (or 'none'):"
scripts/hitl-loop.template.sh:42来自代码打开原文件
printf '\n--- Captured ---\n'printf 'ERRORED=%s\n' "$ERRORED"printf 'ERROR_MSG=%s\n' "$ERROR_MSG"
从这里开始 · 工作说明SKILL.md
diagnosing-bugs
连线表示工作说明包含的模块，不是实际运行顺序。点击模块可查看原文。

文件引用关系图

1 处引用
哪些文件发起引用引用了什么
连线表示真实的文件引用，不是运行顺序。点击节点可高亮相关连线，并查看具体文件和原文位置。虚线表示还有文件需要定位。
文件与检查记录3 个文件

检查范围与遗漏

逐文件查看涉及的内容

下方列出本次涉及的原文范围；纳入检查不代表已查清所有问题。

  • SKILL.md已纳入全文
  • scripts/hitl-loop.template.sh已纳入全文
  • agents/openai.yaml已纳入全文

这份报告只针对上方版本。我们看了拿到的代码和说明文件，没有实际运行 Skill，也没有检查它另外安装的软件包。因此，这不是“保证安全”的承诺；换了版本或使用环境，结果也可能不同。

  • SKILL.md工作说明
  • agents/openai.yaml配套文件
  • scripts/hitl-loop.template.sh脚本

代码和说明中提到的操作

连接外部网站
scripts/hitl-loop.template.sh:34来自代码打开原文件
step "Open the app at http://localhost:3000 and sign in."
SKILL.md:27来自说明文档打开原文件
1. **Failing test** at whatever seam reaches the bug: unit, integration, e2e.2. **Curl / HTTP script** against a running dev server.3. **CLI invocation** with a fixture input, diffing stdout against a known-good snapshot.
SKILL.md:59来自说明文档打开原文件
Phase 1 is done when the loop is **tight** and **red-capable**: you can name **one command** (a script path, a test invocation, a curl) that you have **already run at least once** (show the invocation and its output, redacted), and that is:
运行命令
scripts/hitl-loop.template.sh:1来自代码打开原文件
#!/usr/bin/env bash# Human-in-the-loop reproduction loop.
scripts/hitl-loop.template.sh:7来自代码打开原文件
# Usage:#   bash hitl-loop.template.sh#
SKILL.md:35来自说明文档打开原文件
9. **Differential loop.** Run the same input through old-version vs new-version (or two configs) and diff outputs.10. **HITL bash script.** Last resort. If a human must click, drive _them_ with `scripts/hitl-loop.template.sh` so the loop is still structured. Captured output feeds back to you.
读取了多少行
188
文件校验值（用于核对版本）
8be72a06260e95fe7b5213272703b7d5674782ad710e23a56476e63ed4217c76