DotCraft Hooks Samples

This sample shows how to configure DotCraft lifecycle hooks in your workspace to block dangerous operations, log tool results, and extend agent behavior.

Sample Contents

This sample provides two ready-to-use platform directories:

• windows: Windows / PowerShell example
• linux: Linux / macOS Bash example

Each directory includes:

• hooks.json
• hooks/guard-exec.*: blocks obviously dangerous Exec commands during PreToolUse
• hooks/log-post-tool.*: logs WriteFile / EditFile operations during PostToolUse

Usage

Copy the files from the platform directory you want into your real workspace:

<workspace>/.craft/hooks.json
<workspace>/.craft/hooks/...

If you use the shell scripts on Linux / macOS, remember to make them executable:

chmod +x hooks/*.sh

create-hooks Skill

DotCraft includes a built-in create-hooks skill that can help you generate or adjust hooks.json and hooks scripts during an AI conversation.

If you do not want to write the configuration by hand, you can describe needs such as:

• Blocking specific shell commands
• Appending logs after file writes
• Sending notifications after the agent stops responding

Important Note

After creating or modifying hooks, the user must restart DotCraft manually before the changes take effect.

Events Demonstrated

Event	Purpose
PreToolUse	Checks Exec commands before execution and blocks obviously dangerous ones
PostToolUse	Appends a log entry after a file write or edit succeeds

Expected Behavior

• When the agent tries to run an obviously dangerous shell command, the hook returns a blocking reason
• When the agent successfully runs WriteFile or EditFile, a log entry is appended to hooks/hooks.log