remote-ops

Purpose

Guide and execute remote operations tasks — deployments, infrastructure management, server administration, monitoring setup, and CI/CD pipeline configuration.

When to Use

• User asks about deployment procedures
• User needs to set up or manage infrastructure
• User wants CI/CD pipeline configuration
• User needs monitoring or alerting setup
• User asks about server administration tasks

When NOT to Use

• User asks for application feature development unrelated to operations
• User requests risky production changes without rollback or validation criteria

Inputs

Required:

• Task description — what operation needs to be performed

Optional:

• Target environment (AWS, GCP, Azure, bare metal)
• Current infrastructure state
• Access constraints and security requirements
• Rollback requirements

Process

Step 1: Assess the Task

Understand what operation is needed, the target environment, and any constraints.

Step 2: Plan the Operation

Create a step-by-step runbook with pre-checks, execution steps, validation, and rollback procedures.

Step 3: Execute or Guide

Either execute the operation directly (if the agent has access) or provide a detailed runbook for the user to follow.

Step 4: Validate

Verify the operation succeeded with health checks and monitoring.

Step 5: Document

Record what was done, any issues encountered, and lessons learned.

Output Format

Runbook in Markdown or direct execution with logged output. Filename: runbook-<operation>.md

Quality Checks

• [ ] Pre-checks, execution steps, validation, and rollback are all defined
• [ ] Commands are environment-specific and include safe defaults
• [ ] Success/failure criteria are observable (health checks, metrics, logs)
• [ ] Any destructive action is explicitly called out before execution

Common Skills Used

• common-skills/output-formatting.md — For runbooks and documentation
• common-skills/quality-checklist.md — Validate the operation plan

Edge Cases

• Limited access/no shell: Provide non-invasive diagnostic commands and escalation path
• Production incident context: Prioritize stabilization and evidence capture before optimization
• Unknown environment drift: Add verification steps before and after each change block