Commands & Scripts

commands: {
  commandId: string     // unique identifier
  name: string          // the slash-command name (e.g., "run-tests")
                        // invoked as /run-tests in Claude Code
  description: string   // what this command does
  content: string       // markdown instructions executed when command is invoked
  isEnabled: boolean    // whether this command is selectable
}

commandVersions: {
  commandId: string
  version: number       // incrementing version number
  name: string          // name at time of publish
  description: string   // description at time of publish
  content: string       // full content at time of publish
  status: "active" | "inactive"
  publishedAt: number   // Unix timestamp
  publishedBy: string   // user ID of publisher
}

# Run Tests

Execute the full test suite and report results.

## Steps
1. Run `bun test --run` to execute all unit and integration tests
2. If tests fail, read the error output carefully and identify the root cause
3. Do NOT modify test files to make tests pass - fix the source code instead
4. After all tests pass, run `bun run build` to verify the TypeScript compiles
5. Report a summary: total tests run, passed, failed, and any build warnings

# Security Review

Perform a targeted security audit on all files changed in the current session.

## Checklist
1. **Input validation** - Confirm all user-supplied inputs are validated before use
2. **SQL/command injection** - Verify no user input is interpolated into queries or shell commands
3. **Authentication** - Check that all API routes requiring auth are protected
4. **Secrets** - Scan for hardcoded API keys, passwords, or tokens (use environment variables instead)
5. **XSS** - Confirm no `dangerouslySetInnerHTML` usage without sanitization
6. **Dependencies** - Flag any newly added dependencies with known vulnerabilities

## Output
Return a structured report:
- PASS if no issues found
- FAIL with a numbered list of issues, each with the file path, line number, and recommended fix

# Deploy Check

Verify the application is ready for deployment.

## Pre-deploy Verification Steps
1. Run `bun run build` - must exit with code 0
2. Run `bun run type-check` - zero TypeScript errors permitted
3. Run `bun run lint` - zero ESLint errors permitted (warnings are acceptable)
4. Run `bun test --run` - all tests must pass
5. Check for uncommitted changes with `git status`
6. Confirm the branch is up to date with main using `git log main..HEAD --oneline`

## Output
If all checks pass: respond with "DEPLOY_READY" and a brief summary of what was verified.
If any check fails: respond with "DEPLOY_BLOCKED", the specific check that failed, and the fix required.

scripts: {
  scriptId: string      // unique identifier
  name: string          // display name (e.g., "setup-environment")
  description: string   // what this script does and when to run it
  content: string       // full script content (shell, Python, etc.)
  isEnabled: boolean    // whether this script is selectable
}

scriptVersions: {
  scriptId: string
  version: number       // incrementing version number
  name: string          // name at time of publish
  description: string   // description at time of publish
  content: string       // full script content at time of publish
  status: "active" | "inactive"
  publishedAt: number   // Unix timestamp
  publishedBy: string   // user ID of publisher
}

#!/bin/bash
# Setup Environment
# Run this script at the start of a new session to initialize the dev environment.

set -e  # Exit on any error

echo "==> Cloning repository..."
cd /home/user
git clone "$REPO_URL" project
cd project

echo "==> Installing dependencies..."
bun install

echo "==> Setting up environment variables..."
cp .env.example .env.local
echo "NEXT_PUBLIC_CONVEX_URL=$CONVEX_URL" >> .env.local
echo "NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=$CLERK_KEY" >> .env.local

echo "==> Running initial build to verify setup..."
bun run build

echo "==> Environment ready. Start dev server with: bun dev"

#!/bin/bash
# Run End-to-End Tests
# Starts the dev server, waits for readiness, then runs Playwright tests.

set -e

echo "==> Starting dev server in background..."
nohup bun dev > /tmp/devserver.log 2>&1 &
DEV_PID=$!

echo "==> Waiting for server to be ready..."
timeout 60 bash -c 'until curl -sf http://localhost:3000 > /dev/null; do sleep 2; done'

echo "==> Running Playwright E2E tests..."
bun run test:e2e

echo "==> Stopping dev server..."
kill $DEV_PID 2>/dev/null || true