automation-helper

name: automation-helper description: Design and create automation workflows — shell scripts, Cursor hooks, GitHub Actions, cron jobs, and task runners. Use when the user asks to automate a process, create a script, or set up a workflow.

Skill: Automation Helper

Design and implement automation workflows for repetitive development tasks using shell scripts, Cursor hooks, GitHub Actions, or task runners.

Trigger

When the user asks to automate a process, create a shell script, set up a GitHub Action, add a Cursor hook, or streamline a repetitive workflow.

Prerequisites

• Clear understanding of the task to automate
• Target platform identified (local, CI, Cursor hook, cron)

Steps

Step 1: Identify Automation Type

Step 2: Gather Requirements

• What triggers the automation? (manual, git event, schedule, file change)
• What inputs does it need? (files, env vars, user input)
• What does it produce? (files, logs, notifications, side effects)
• What are the failure modes? (missing deps, network errors, permissions)
• Should it be idempotent? (safe to re-run)

Step 3: Design the Workflow

• Break into discrete steps (each step should do one thing)
• Identify dependencies between steps
• Define success/failure criteria for each step
• Plan error handling and rollback
• Determine logging and notification needs

Step 4: Implement

Shell Script Template

#!/usr/bin/env bash
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"

log() { echo "[$(date '+%Y-%m-%d %H:%M:%S')] $*"; }
error() { log "ERROR: $*" >&2; exit 1; }

# Validate prerequisites
command -v node >/dev/null 2>&1 || error "node is required"

log "Starting [task name]..."

# Step 1: [description]
log "Step 1: [description]"
# implementation here

# Step 2: [description]
log "Step 2: [description]"
# implementation here

log "Done."

GitHub Actions Template

name: [Workflow Name]
on:
  [trigger]:
    branches: [main]
    paths:
      - 'src/**'

jobs:
  [job-name]:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with:
          node-version: '20'
          cache: 'npm'
      - run: npm ci
      - run: [command]

Cursor Hook Template (.cursor/hooks/)

#!/usr/bin/env bash
set -euo pipefail

# Hook: [name]
# Trigger: [when this runs]
# Purpose: [what it does]

[implementation]

Makefile Template

.PHONY: [target]

[target]: ## [description]
    @echo "Running [target]..."
    [command]

Step 5: Add Error Handling

• Use set -euo pipefail in all bash scripts
• Add meaningful error messages with context
• Implement cleanup on failure (trap EXIT)
• Add retry logic for transient failures (network, rate limits)
• Log failures with enough context to debug

cleanup() {
  local exit_code=$?
  if [ $exit_code -ne 0 ]; then
    log "Script failed with exit code $exit_code"
    # cleanup actions here
  fi
}
trap cleanup EXIT

Step 6: Add Documentation

• Header comment with purpose, trigger, prerequisites
• Usage examples in the script or README
• Document required environment variables
• Document expected inputs and outputs

Step 7: Test the Automation

• Run manually with expected inputs
• Test failure scenarios (missing inputs, network errors)
• Test idempotency (run twice — same result)
• Verify logging output is useful
• Verify no secrets are logged or exposed

Step 8: Wire Up the Trigger

• For hooks: add to .cursor/hooks.json
• For CI: add to .github/workflows/ or .gitlab-ci.yml
• For cron: add crontab entry or CloudWatch rule
• For manual: add to Makefile or package.json scripts

Rules

• ALWAYS use set -euo pipefail in bash scripts
• ALWAYS validate inputs and prerequisites before executing
• NEVER hardcode secrets — use environment variables
• NEVER use rm -rf / or other dangerous commands without safeguards
• ALWAYS add a --dry-run option for destructive operations
• Scripts must be idempotent unless explicitly documented otherwise
• Prefer #!/usr/bin/env bash over #!/bin/bash for portability

Common Automations

Completion

Working automation with error handling, documentation, and trigger wired up. Tested manually at minimum.

If a Step Fails

• Missing command: Check prerequisites and add installation step
• Permission denied: Verify chmod +x on scripts and IAM/role permissions for CI
• Flaky in CI: Add retries, increase timeouts, check for race conditions
• Too slow: Parallelize independent steps, cache dependencies