Automation Recipes

Create a pipeline called "post-deploy-checks":
1. Trigger (webhook, path: "/hooks/deploy-complete")
2. HTTP Request — GET https://api.yourapp.com/health
3. Condition — response status is 200
4. True: Agent (tester) — "Run smoke tests against production: check auth flow, main API endpoints, and critical user paths"
5. False: Skill (telegram-notify) — "CRITICAL: Deploy health check failed!"
6. Memory Store — store deploy result with timestamp

Create a pipeline called "dependency-scanner":
1. Trigger (schedule, cron: "0 10 * * 1") — every Monday at 10am
2. Agent (analyst) — "Analyze package.json for outdated dependencies and known vulnerabilities. Check npm audit results."
3. Condition — vulnerabilities found?
4. True: Create kanban task with P1 priority + Skill (slack-notify)
5. False: Output (log) — "All dependencies up to date"

Create a pipeline called "release-notes":
1. Trigger (manual) — version number as payload
2. Agent (researcher) — "Gather all completed kanban tasks since last release"
3. Agent (developer) — "Generate release notes from these tasks. Group by: features, fixes, improvements"
4. Agent (reviewer) — "Review release notes for clarity and completeness"
5. Output (file) — save to CHANGELOG.md
6. Skill (slack-notify) — announce release in #releases

Create a pipeline called "multi-health-check":
1. Trigger (schedule, cron: "*/10 * * * *") — every 10 minutes
2. Split — split list of endpoints: ["/api/health", "/api/auth/status", "/api/billing/status"]
3. Loop — for each endpoint:
   a. HTTP Request — GET the endpoint
   b. Code — check response time and status
4. Aggregate — collect all results
5. Filter — keep only failures
6. Condition — any failures?
7. True: Agent (analyst) — "Analyze these failures and suggest root cause: {{input}}"
   → Skill (telegram-notify) — alert with analysis
8. Memory Store — store health check results for trend analysis

Create a pipeline called "error-rate-monitor":
1. Trigger (schedule, cron: "0 * * * *") — every hour
2. HTTP Request — query your logging service for error count
3. Memory Search — find last hour's error count from memory
4. Code — calculate error rate change percentage
5. Condition — rate increase > 50%?
6. True: Agent (analyst) — "Error rate spiked. Analyze recent errors and suggest fixes"
   → Skill (telegram-notify) — alert the team
7. Memory Store — store current error count for next comparison

Create a pipeline called "docs-freshness":
1. Trigger (schedule, cron: "0 9 * * 1") — every Monday
2. Agent (researcher) — "Compare documentation in /docs with current codebase. Find any outdated references, missing new features, or incorrect examples"
3. Condition — outdated docs found?
4. True: Create kanban tasks for each outdated section
5. Output (log) — freshness report

Create a pipeline called "meeting-processor":
1. Trigger (manual) — paste meeting notes
2. Agent (analyst) — "Extract action items, decisions, and follow-ups from these meeting notes"
3. Agent (manager) — "Create kanban tasks from the action items with appropriate priorities and assignees"
4. Memory Store — store key decisions as architecture/service-context memories
5. Skill (slack-notify) — post action items summary to #team

Create a pipeline called "test-generator":
1. Trigger (webhook, path: "/hooks/new-code")
2. Memory Search — find testing patterns and conventions
3. Agent (tester) — "Generate comprehensive tests for this code following our testing conventions: {{input}}"
4. Agent (reviewer) — "Review the generated tests for coverage and correctness"
5. Output (file) — save tests to appropriate __tests__ directory

Create a pipeline called "security-audit":
1. Trigger (schedule, cron: "0 10 * * 5") — every Friday
2. Agent (analyst) — "Perform a security audit: check for hardcoded secrets, SQL injection risks, XSS vulnerabilities, and insecure dependencies"
3. Memory Search — find previous audit results for comparison
4. Agent (reviewer) — "Compare with previous audit and highlight new issues"
5. Condition — critical issues found?
6. True: Create P0 kanban task + Skill (telegram-notify)
7. Memory Store — store audit results