Writing Hookify Rules

Hookify rules enforce coding standards using YAML frontmatter (name, enabled, event, pattern, conditions) and regex patterns to match specific code, commands, or prompts. Define the event trigger (bash, file, prompt, or stop), set your pattern, and configure the rule to warn or block non-compliant actions, outputting a Markdown document with your rule definition and message body.

Hookify rules trigger on four event types: bash commands, file modifications, user prompt submissions, and agent stop events. Each event type activates pattern matching against the specified regex, allowing you to intercept and respond to different workflow stages with warnings or blocks.

Yes, Hookify rules use regex patterns to match specific commands, file edits, or user prompts. Pattern-based matching lets you target dangerous operations (like rm -rf), unsafe practices (console.log in production), or custom security requirements, then warn or block matches automatically.

Hookify rules support two actions: warn and block. Set the action in your rule configuration to warn Claude about potential issues while allowing the operation, or block to prevent the operation entirely. This lets you guide behavior from advisory feedback to hard enforcement.

The Hookify rule schema includes name (rule identifier), enabled (boolean), event (bash, file, prompt, or stop), pattern (regex expression), and conditions (filter criteria). Combine these frontmatter fields with your rule message body in a Markdown document to define enforcement logic.

Yes, Hookify rules can block or warn on dangerous bash commands using regex patterns. Define a rule with event set to bash, pattern matching the command (like sudo), and action set to block or warn, preventing unsafe operations or alerting before execution.