Evaluate hookify plugin to replace bespoke hook scripts

[philoserf]

Problem

config-protection.sh, validate-bash-commands.py, and parts of prompt-injection-guard.py each implement pattern-matching-and-warn logic in bespoke scripts. Adding a new rule means editing Python or bash code.

Solution

The official hookify plugin (claude-code/plugins/hookify/) provides a declarative rule engine where rules are markdown files with frontmatter:

---
name: block-dangerous-rm
event: bash
pattern: rm\s+-rf
action: block
---
⚠️ Dangerous rm command detected!

Features:

• Regex matching, substring contains, equals, starts_with, ends_with operators
• Multi-condition rules (all conditions must match)
• Block vs. warn actions
• Field extraction for Bash commands, file paths, edit content
• Works on PreToolUse, PostToolUse, Stop, UserPromptSubmit events

Could replace 3 hook scripts with ~5 declarative .local.md rule files — less code to maintain.

Trade-offs

• Adds a plugin dependency vs. self-contained scripts
• May not cover all current hook logic (e.g., prompt-injection-guard's invisible unicode detection)
• Need to verify it handles the config-protection case statement patterns