Skip to main content

In this article

Stage 9: Operations

Overview

Operations covers the ongoing lifecycle after delivery, including incident response, documentation maintenance, prompt refinement, and system monitoring. This stage provides tooling for keeping production systems healthy and documentation current.

When You Enter This Stage

You enter Operations after completing the final sprint delivery in Stage 8: Delivery.

NOTE

Prerequisites: Production deployment complete. Monitoring and alerting configured.

Available Tools

Primary Agents

ToolTypeHow to InvokePurpose
documentationAgentSelect documentation agentAudit, drift, author, and validate documentation

Prompts

ToolTypeHow to InvokePurpose
incident-responsePrompt/incident-responseDocument and triage incidents

Auto-Activated Instructions

InstructionActivates OnPurpose
writing-style**/*.mdEnforces voice and tone conventions
markdown**/*.mdEnforces Markdown formatting rules
hve-builderAI artifactsEnforces authoring standards

Skills

SkillPurpose
hve-builderReview, refactor, or validate operational AI artifacts
rpi-researchInvestigate decision-critical operational evidence gaps

Templates

TemplatePurpose
incident-responseStructured template for incident documentation

Role-Specific Guidance

SREs lead Operations, handling incident response and system monitoring. Tech Leads contribute to architecture-level maintenance decisions. Engineers address hotfixes and ongoing code maintenance.

Starter Prompts

Incident Response

/incident-response Users are reporting 504 Gateway Timeout errors on the
/api/v2/orders endpoint in East US 2. Errors started at 14:32 UTC after
the App Service scaled down during a scheduled maintenance window.
Severity 2. Phase is triage.

Documentation Maintenance

Select the documentation agent and choose the validate or author mode to target a specific scope and focus area:

Select documentation agent in validate mode. Scope docs, focus accuracy.

For ad-hoc documentation work, select the agent and describe the task directly.

Select documentation agent:

Scan docs/getting-started/ for accuracy against current scripts/ and
.github/ artifacts. Several install scripts changed in the v3.2 release
and the setup guides may reference outdated flags or file paths.

Prompt Refinement

Use hve-builder with mode=review and
targets=.github/prompts/security/incident-response.prompt.md. Evaluate its
activation, operational safeguards, output contract, and host compatibility.

After review, use hve-builder improve or refactor mode only when source changes are approved.

To create a new prompt from an existing implementation file, use hve-builder create mode:

Use hve-builder with mode=create. Create a prompt from
src/api/handlers/search.py that generates search
handler implementations following the same query parsing, pagination,
and response envelope patterns.

Operational Continuity

Resume operational work from the artifacts owned by the active workflow. For incident response, reopen the current incident report or runbook and continue from its recorded status and next actions. For documentation work, resume from the Documentation workflow's session record and target files. For HVE Builder, resume from its author, review, behavior, and validation evidence.

Stage Outputs and Next Stage

Operations produces updated documentation, incident reports, refined prompts, and maintenance artifacts. When a hotfix is needed, transition back to Stage 6: Implementation to address the issue through the standard implementation workflow.

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.