feat(j10c-frontend-playbook): add fix-patterns from WeJH-Taro audit

[j10ccc]

Summary

Adds skills/j10c-frontend-playbook/fix-patterns/ — an empirical weakness pattern library derived from a full-repo audit of zjutjh/WeJH-Taro (321 fix commits, 2023-08 to 2026-05).

Files added

File	Content
fix-patterns/INDEX.md	Pattern index with commit counts, percentages, and links
fix-patterns/weak-ui-style.md	CSS isolation / z-index / dark mode (101 commits, 31.5%)
fix-patterns/weak-logic-bug.md	Float precision / key uniqueness / boundary guards (89 commits, 27.7%)
fix-patterns/weak-type-null.md	Type guards / null-undefined safety (34 commits, 10.6%)

Each pattern file includes:

• One-line definition
• Trigger conditions
• 3 real commit references with GitHub URLs
• Self-check list
• Fix template (code snippet)

Data source

Full audit pipeline:

1. gh repo clone zjutjh/WeJH-Taro → extracted all 321 fix commits via git log --grep=^fix
2. For each commit: git show --stat + diff preview → auto-classified into a 3-level taxonomy
3. Weakness analysis → 3 top weak areas identified
4. Each weakness → structured pattern doc

Output artifacts at ~/Desktop/wejh-fix-audit/:

• commits.md — full audit table (hash / date / author / scope / one-liner / category)
• taxonomy.md — classification tree (14 leaf nodes, 3 levels)
• analysis.md — count table + mermaid pie + weakness summary + self-check list

Continuous monitoring

A Hermes cron job (daily-pr-fix-pattern-review, ID: e10817b586fb) runs at 23:00 daily:

• Scans gh pr list --author j10ccc --state merged for the last 24h
• Analyzes each PR diff against known weak patterns
• Writes daily report to ~/Desktop/wejh-fix-audit/daily-pr-review/<YYYY-MM-DD>.md
• Appends matched examples to the corresponding pattern file