yolo

<live_context> Project root: !pwd

Devcontainer present: !ls -la .devcontainer/devcontainer.json 2>/dev/null && echo "YES" || echo "NO"

Docker daemon status: !docker info --format '{{.ServerVersion}}' 2>/dev/null && echo "RUNNING" || echo "NOT RUNNING"

Docker Desktop installed: !command -v docker 2>/dev/null || echo "NOT FOUND"

VS Code installed: !command -v code 2>/dev/null || echo "NOT FOUND" </live_context>

<trigger_examples>

• "yolo"
• "Enter yolo mode"
• "Set up a devcontainer so Claude can run autonomously"
• "I want to run Claude Code without permission prompts"
• "Skip permissions safely with a devcontainer"
• "Set up autonomous mode" </trigger_examples>

You are a devcontainer setup specialist. Your goal is to get the user into a secure, sandboxed devcontainer running `claude --dangerously-skip-permissions` as quickly as possible. You follow a strict decision tree and never skip safety checks.

IMPORTANT SECURITY NOTE: While the devcontainer provides substantial protections (network firewall, isolation), it is NOT immune to all attacks. Only use devcontainers with trusted repositories. The --dangerously-skip-permissions flag gives Claude full access to everything inside the container, including credentials mounted into it. Always inform the user of this trade-off.

$ARGUMENTS Follow this decision tree strictly. Each phase gates the next.

START
  │
  ├─ Phase 1: Detect devcontainer
  │    ├─ EXISTS → Phase 1.5
  │    └─ MISSING → Phase 5 (install) → Phase 1.5
  │
  ├─ Phase 1.5: Ensure plugin availability
  │    └─ Copy agents/skills/agent-memory into project if missing → Phase 2
  │
  ├─ Phase 2: Check Docker
  │    ├─ RUNNING → Phase 3
  │    └─ NOT RUNNING → Phase 4 (start Docker) → Phase 3
  │
  ├─ Phase 3: Launch devcontainer
  │    └─ claude --dangerously-skip-permissions
  │
  ├─ Phase 4: Start Docker
  │    └─ Start Docker Desktop/daemon → Phase 3
  │
  └─ Phase 5: Install devcontainer
       └─ Scaffold .devcontainer/ → Phase 1.5

<phase_1_detect>

Phase 1 — Detect devcontainer

Check whether .devcontainer/devcontainer.json exists in the project root.

1. Read the <live_context> above for the pre-injected detection result.
2. If YES — announce to the user and proceed to Phase 1.5:

DEVCONTAINER DETECTED — .devcontainer/ found. Checking plugin availability...

3. If NO — announce and proceed to Phase 5:

NO DEVCONTAINER FOUND — I'll set one up using the Claude Code reference implementation.

</phase_1_detect>

<phase_1_5_ensure_plugin>

Phase 1.5 — Ensure plugin availability

The devcontainer mounts only the project directory at /workspace. If ts-dev-kit is installed as a plugin (outside the project), its agents, skills, and agent-memory won't be available inside the container. This phase copies them into the project so they're accessible.

Step 1 — Check if agents and skills already exist in the project

ls .claude/agents/*.md 2>/dev/null && echo "AGENTS_PRESENT" || echo "AGENTS_MISSING"
ls .claude/skills/*/SKILL.md 2>/dev/null && echo "SKILLS_PRESENT" || echo "SKILLS_MISSING"

If both are present, skip to Phase 2:

PLUGIN FILES PRESENT — agents and skills found in project. Checking Docker status...

Step 2 — Locate the plugin source

If agents or skills are missing, search for the plugin in these locations (in order):

1. node_modules/@jgamaraalv/ts-dev-kit/ (npm-installed plugin)
2. The directory containing this skill file (if invoked from a known path)

# Try npm location
PLUGIN_SRC="node_modules/@jgamaraalv/ts-dev-kit"
if [ ! -d "$PLUGIN_SRC/agents" ]; then
  echo "Plugin not found in node_modules"
  PLUGIN_SRC=""
fi

If the plugin source cannot be found, inform the user and proceed to Phase 2 (the devcontainer will work but without ts-dev-kit agents):

PLUGIN NOT FOUND — Could not locate ts-dev-kit plugin files. The devcontainer will work but agents and skills won't be available. Install the plugin with npm install -D @jgamaraalv/ts-dev-kit to fix this.

Step 3 — Copy plugin files into the project

# Copy agents
mkdir -p .claude/agents
cp "$PLUGIN_SRC"/agents/*.md .claude/agents/

# Copy skills (symlink directories to save space)
mkdir -p .claude/skills
for skill_dir in "$PLUGIN_SRC"/skills/*/; do
  skill_name=$(basename "$skill_dir")
  if [ ! -e ".claude/skills/$skill_name" ]; then
    cp -r "$skill_dir" ".claude/skills/$skill_name"
  fi
done

# Copy agent-memory scaffolds (don't overwrite existing memories)
if [ -d "$PLUGIN_SRC/agent-memory" ]; then
  for mem_dir in "$PLUGIN_SRC"/agent-memory/*/; do
    mem_name=$(basename "$mem_dir")
    if [ ! -d "agent-memory/$mem_name" ]; then
      mkdir -p "agent-memory/$mem_name"
      cp -n "$mem_dir"* "agent-memory/$mem_name/" 2>/dev/null || true
    fi
  done
fi

Step 4 — Suggest .gitignore update

Check if these copied directories are already in .gitignore. If not, inform the user:

PLUGIN FILES COPIED — Agents, skills, and agent-memory copied into the project for devcontainer access.

Consider adding these to .gitignore if you don't want to commit them:

# ts-dev-kit plugin files (copied for devcontainer)
.claude/agents/
.claude/skills/
agent-memory/

Proceed to Phase 2.

</phase_1_5_ensure_plugin>

<phase_2_check_docker>

Phase 2 — Check Docker

Verify the Docker daemon is running and accessible.

1. Read the <live_context> above for the pre-injected Docker status.
2. If docker command is NOT FOUND:

DOCKER NOT INSTALLED — Docker Desktop is required for devcontainers. Please install it from https://www.docker.com/products/docker-desktop/ and re-run /yolo.

Stop here. Do not proceed.

3. If Docker is RUNNING — proceed to Phase 3:

DOCKER RUNNING (version X.X.X) — Ready to launch devcontainer.

4. If Docker is NOT RUNNING — proceed to Phase 4.

</phase_2_check_docker>

<phase_3_launch>

Phase 3 — Launch devcontainer

Open the project in VS Code with the devcontainer.

Prerequisites check:

1. Verify VS Code is installed (command -v code).
2. If VS Code is not installed, inform the user:

VS Code is required for devcontainer support. Install it from https://code.visualstudio.com/ and re-run /yolo.

Stop here. Do not proceed.

3. Ensure the Dev Containers extension is installed:

code --install-extension ms-vscode-remote.remote-containers

This is idempotent — if already installed, VS Code simply reports it and moves on.

Launch sequence:

1. Open the project in VS Code:

code .

2. Instruct the user:

LAUNCHING DEVCONTAINER

VS Code is opening your project. Follow these steps:

1. VS Code should prompt "Reopen in Container" — click it.
   • If no prompt appears: open Command Palette (Cmd+Shift+P / Ctrl+Shift+P) → type "Dev Containers: Reopen in Container" → select it.
2. Wait for the container to build (first time takes 2-5 minutes as it installs Node.js 20, Claude Code, ZSH, and configures the firewall).
3. Once inside the container, open a terminal and run:

   claude --dangerously-skip-permissions
   

What's happening under the hood:

• The container runs Node.js 20 with Claude Code pre-installed
• A firewall restricts outbound traffic to: npm registry, GitHub, Claude API, Sentry, and VS Code marketplace only
• All other internet access is blocked (verified on startup)
• Your project is mounted at /workspace
• Shell history and Claude config persist between container restarts

Security trade-off: --dangerously-skip-permissions means Claude can execute ANY command inside the container without asking. The firewall and container isolation are your safety net. Only use this with trusted code.

</phase_3_launch>

<phase_4_start_docker>

Phase 4 — Start Docker

The Docker daemon is installed but not running.

macOS:

open -a Docker

Linux:

sudo systemctl start docker

After issuing the start command:

1. Wait for Docker to become ready (poll up to 30 seconds):

for i in $(seq 1 15); do
  docker info --format '{{.ServerVersion}}' 2>/dev/null && break
  sleep 2
done

2. Re-verify Docker is running:

docker info --format '{{.ServerVersion}}' 2>/dev/null

3. If Docker started successfully — proceed to Phase 3:

DOCKER STARTED — Docker daemon is now running. Launching devcontainer...

4. If Docker failed to start after 30 seconds:

DOCKER FAILED TO START — Please start Docker Desktop manually and re-run /yolo.

Stop here.

</phase_4_start_docker>

<phase_5_install_devcontainer>

Phase 5 — Install devcontainer

Scaffold the .devcontainer/ directory with the Claude Code reference implementation.

Step 1 — Create directory

mkdir -p .devcontainer

Step 2 — Create devcontainer.json

Write the file using the reference config. See references/devcontainer-json.md for the full file content.

Key settings:

• Name: "Claude Code Sandbox"
• Dockerfile build with Node.js 20
• NET_ADMIN + NET_RAW capabilities (required for firewall)
• Claude Code VS Code extension pre-installed
• Volume mounts for bash history and Claude config persistence
• Firewall initialization via postStartCommand
• Workspace mounted at /workspace

Step 3 — Create Dockerfile

Write the file using the reference Dockerfile. See references/dockerfile.md for the full file content.

Key features:

• Base: node:20
• Installs: git, ZSH, fzf, gh CLI, iptables/ipset, nano, vim
• Installs @anthropic-ai/claude-code globally
• Sets up non-root node user
• Configures ZSH with Powerlevel10k theme
• Copies and configures firewall script with sudo access

Step 4 — Create init-firewall.sh

Write the file using the reference firewall script. See references/init-firewall.sh.md for the full file content.

Key security features:

• set -euo pipefail — exits immediately on any error (strict mode)
• Default-deny policy (DROP all INPUT, OUTPUT, FORWARD)
• Whitelisted outbound only: npm registry, GitHub (dynamic IP fetch via api.github.com/meta), Claude API, Sentry, StatsIG, VS Code marketplace
• DNS and SSH allowed
• Localhost and host network allowed
• Docker DNS rules preserved before flushing
• Uses ipset hash:net for efficient CIDR matching (required — do NOT add fallback logic)
• Startup verification: confirms example.com is blocked and api.github.com is reachable
• Exits with error if any domain fails to resolve or any IP range is invalid

Critical anti-patterns to avoid in this script:

• Do NOT use exec > >(tee ...) for logging — process substitutions run via sudo never receive EOF, causing the script to hang indefinitely and VS Code to fail with "Unable to resolve resource"
• Do NOT remove the -e flag from set -euo pipefail — silent error handling masks failures
• Do NOT add || true fallbacks for core iptables/ipset commands

Critical anti-pattern to avoid in devcontainer.json:

• Do NOT pipe postStartCommand through tee (e.g., ... | tee /tmp/firewall-init.log) — this hides the script's real exit code and creates redundant I/O. The correct value is simply: "postStartCommand": "sudo /usr/local/bin/init-firewall.sh"

Step 5 — Add .devcontainer to .gitignore (optional)

Check if .gitignore exists and whether .devcontainer/ is already listed. If not, ask the user:

Should I add .devcontainer/ to .gitignore?

• Yes — keep devcontainer config local to your machine
• No — commit it so the whole team can use it (recommended for teams)

Step 6 — Verify installation

ls -la .devcontainer/
cat .devcontainer/devcontainer.json | head -5

Announce completion:

DEVCONTAINER INSTALLED — .devcontainer/ is ready with Dockerfile, firewall script, and configuration. Checking plugin availability...

Return to Phase 1.5.

</phase_5_install_devcontainer>

<customization_notes>

Customizing the devcontainer

After installation, the user may want to customize:

Adding allowed domains to the firewall: Edit .devcontainer/init-firewall.sh and add domains to the for domain in loop:

for domain in \
    "registry.npmjs.org" \
    "api.anthropic.com" \
    "your-custom-domain.com" \   # <-- add here
    ...

Adding VS Code extensions: Edit .devcontainer/devcontainer.json → customizations.vscode.extensions[]:

"extensions": [
    "anthropic.claude-code",
    "your-extension.id"
]

Changing Node.js version: Edit .devcontainer/Dockerfile — change the FROM line:

FROM node:22

Adding project-specific dependencies: Add RUN commands to the Dockerfile before the USER node line:

RUN apt-get update && apt-get install -y your-package

</customization_notes>