What problem does it solve?
Debugging issues involving browsers, editors, CDN/WAF layers, IM platforms, OS clipboards, or third-party SaaS often fails because the behavior lives outside your code. This Skill stops the cycle of repeated code-reasoning fixes that do nothing by forcing real environment data capture first.
Core Features & Use Cases
- Black-box identification checklist: Determines whether a system qualifies as a black box based on control, documentation, debuggability, and predictability criteria.
- Data collection templates: Provides per-system capture recipes, such as reading all clipboard MIME types via navigator.clipboard.read(), diffing outerHTML before and after paste, and using curl -v to inspect CDN/WAF headers.
- Case library with naming conventions: Ships real cases (CSDN paste list bug, HTTP CDN/WAF interception) plus templates for documenting new findings in references/.
- Use Case: When a copied article renders as four separate "1." list items in CSDN but works in WeChat, use this Skill to capture the actual DOM structure, identify the undocumented HTML-to-Markdown start-attribute loss, and apply a verified single-variable fix.
Quick Start
Ask the AI to debug a paste or third-party API issue using the external-system-debugging skill and have it capture real environment data before proposing any fix.