Why is exit code 2 important in hooks?

In Claude Code, exit code 2 is required to trigger a blocking error that stops an action and shows an error to the user, while exit code 0 signifies success.

What are Claude Code hooks?

Hooks are automated actions triggered at specific points in a Claude Code session, enabling validation, monitoring, and control through bash commands or LLM-based evaluation.

Which events support prompt-type hooks?

Prompt-based hooks, which use an LLM to make decisions, are currently only supported for 'Stop' and 'SubagentStop' events.

How does this skill improve hook safety?

It detects dangerous command patterns like rm -rf or curl | sh and ensures scripts use 'set -euo pipefail' to prevent silent failures.

Hook Security & Enhancement

Name: Hook Security & Enhancement
Author: agent-sh

byagent-sh

•

692

•

セキュリティとテスト

Analyzes and optimizes Claude Code hooks for safety, proper lifecycle execution, and standardized configuration.

This skill acts as a specialized auditor for Claude Code hook implementations, ensuring that automation scripts are safe, performant, and correctly integrated into the session lifecycle. It identifies dangerous bash patterns such as recursive deletes or remote execution, validates exit codes for blocking logic, and ensures prompt-based hooks are used only on supported events. By providing automated reporting and one-click fixes for common configuration issues, it helps developers maintain robust and secure AI-driven workflows.

主な機能

01Validation of hook lifecycle events and matcher syntax

02Verification of exit codes (0 for success, 2 for blocking)

03Automated safety header injection (set -euo pipefail)

04Timeout configuration audits for network-dependent hooks

05692 GitHub stars

06Security auditing for dangerous bash patterns like rm -rf and eval

ユースケース

01Standardizing hook configurations and frontmatter across large-scale projects

02Auditing custom security firewalls to ensure they correctly block malicious tool usage

03Debugging hook failures by verifying event and type compatibility

主な機能

01Validation of hook lifecycle events and matcher syntax

02Verification of exit codes (0 for success, 2 for blocking)

03Automated safety header injection (set -euo pipefail)

04Timeout configuration audits for network-dependent hooks

05692 GitHub stars

06Security auditing for dangerous bash patterns like rm -rf and eval

ユースケース

01Standardizing hook configurations and frontmatter across large-scale projects

02Auditing custom security firewalls to ensure they correctly block malicious tool usage

03Debugging hook failures by verifying event and type compatibility