Docker Sandbox — Security agent for Claude Code
Isolated autonomous scanning agent — runs in a temporary worktree with restricted Bash access.
How to install Docker Sandbox
Installs to ~/.claude/agents/andrem-sec-psc-comet-docker-sandbox.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/andrem-sec/psc-comet/HEAD/.claude/agents/docker-sandbox.md -o ~/.claude/agents/andrem-sec-psc-comet-docker-sandbox.md Restart Claude Code, or start a new session, for it to be picked up.
What Docker Sandbox does
name: docker-sandbox memory_scope: project description: Isolated autonomous scanning agent — runs in a temporary worktree with restricted Bash access. Requires ROE authorization before any execution. Use for security scanning, vulnerability assessment, and autonomous probing of in-scope systems. tools:
- Bash
- Read
- Glob
- Grep model: claude-sonnet-4-6 permissionMode: dontAsk isolation: worktree maxTurns: 30
Docker Sandbox Agent
You are an isolated security scanning agent.
Alternatives in Security
- Gitnexus Security Boundary Reviewer — GitNexus security and trust-boundary reviewer 45.8k ★
- Deps Audit — You are a dependency security expert specializing in vulnerability scanning, license compliance, and 31.9k ★
- Claude Code System Prompts — by Piebald AI - All parts of Claude Code's system prompt, including builtin tool descriptions, sub agent promp 6.3k ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Code Security Reviewer
Performs structured code security and quality audits with restricted tool access. Supports tiered review (T1 p
Container Attacker
Container and Kubernetes security specialist. Handles Docker escape techniques, Kubernetes RBAC abuse, service
Go Security Release
Security scanning, vulnerability assessment, release readiness checks
Hcs Security Reviewer
Independent read-only security review for HCS changes. Checks secrets, sandbox boundaries, audit integrity, fo
Self Audit
Post-loop, proposes-only self-audit — mines objective cost/perf/correctness signals via ${CLAUDE_PLUGIN_ROOT}/
Worktree Agent Security
Implementer agent specialized in security/hardening, working in ITS OWN git worktree. Launched by the orchestr
Related Skills
Enclave
Sandbox for running AI coding agents autonomously: isolated, network-restricted, host-safe
Iterare LLM
Automated, isolated execution of Claude Code prompts in firewall-restricted Docker containers.
Claude Security Research Skill
AI-powered security research assistant for Claude Code — structured assessment workflows, tool orchestration,