dwarvesf

Claude Guardrails — Security skill for Claude Code

Security community

Hardened security configuration for Claude Code; permission deny rules, shell hooks, and prompt injection defense in full and lite variants.

How to install Claude Guardrails

This entry records only its repository, not the path inside it, so there is no exact command to give. Open dwarvesf/claude-guardrails and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Guardrails does

Hardened security configuration for Claude Code; permission deny rules, shell hooks, and prompt injection defense in full and lite variants.

Alternatives in Security

Defense In Depth — Implement multi-layered testing and security best practices 98.1k ★
Security Guardian — Comprehensive security analysis for RTK CLI tool, focusing on command injection, shell escaping, hook security 11.9k ★
Free Code — The free build of Claude Code 8.1k ★

README

claude-guardrails

Hardened security configuration for [Claude Code](https://claude.ai/code) — deny rules, hooks, and prompt injection defense out of the box.

Why This Exists

Claude Code can read your filesystem, run shell commands, and fetch URLs autonomously. A poisoned file in a cloned repo can hijack its behavior via prompt injection. A careless tool call can leak your SSH keys or `.env` secrets. These configs add defense-in-depth so you don't have to think about it on every session.

Full vs Lite

	Lite	Full
Use when	Internal/trusted projects	Open source repos, untrusted codebases, production credentials
Credential deny rules	21 rules (SSH, AWS, GPG, kube, Azure, .env, .pem, destructive Bash, etc.)	40 rules (adds secrets dirs, shell profiles, crypto wallets, etc.)
PreToolUse hooks	4 (destructive deletes, direct push, pipe-to-shell, commit-time secret scan)	6 (adds data exfiltration, permission escalation)
UserPromptSubmit inbound secret scanner	Yes (`scan-secrets.sh`)	Yes (`scan-secrets.sh`)
PreToolUse commit-time secret scan	Yes (`scan-commit.sh`)	Yes (`scan-commit.sh`)
PostToolUse prompt injection scanner	No	Yes (`prompt-injection-defender.sh`)
Privacy env flags	Yes (telemetry, error reports, feedback survey off)	Yes
CLAUDE.md security rules	Yes	Yes
Sandbox guidance	Yes — this is the enforcement layer	Yes + devcontainer pointer
Prereqs	`jq`	`jq`

Quick Start

# Install jq if you don't have it
brew install jq          # macOS
# sudo apt install jq    # Debian/Ubuntu

# Lite (4 hooks, 21 deny rules — for trusted projects)
npx claude-guardrails install

# Full (6 hooks + prompt injection scanner — for untrusted codebases)
npx claude-guardrails install full

The script merges into your existing `~/.claude/settings.json` (backing it up first) and is safe to run repeatedly.

Install from source (git clone)<