Automated Ticket Management and Issues-as-Code

Overview

• Project: Operational Tool Stack
• Status: Active — evaluating Issues-as-Code principles
• Problem: Stagnant Linear backlog caused by impedance gap between planning tools and execution environment

Open Source Ticket Systems Evaluation

Candidate	Stack	Best For	IaC Compatibility
FreeScout	PHP (Laravel)	Lightweight customer support	Moderate — excellent API, “Help Scout” feel
Peppermint	Node.js/Vue	Modern IT management	High — dev-friendly modern standards
UVdesk	PHP (Symfony)	E-commerce/multi-channel	Low — merchant-customer focused
Zammad	Ruby on Rails	Enterprise scaling	Moderate — strong REST API but heavy

Dark Horse: Local Markdown-based tracking (.imd format) synced to GitHub/GitLab — bypasses GUI helpdesk entirely.

Issues-as-Code Principles

1. Versioned State — issues stored as flat files in docs/tickets/ or .github/issues/
2. Git-Based Triage — status changes via branch creation or file metadata edits
3. Single Source of Truth — IDE is the primary interface; leaving VS Code to update status = system failure

AI-Powered Automation Strategy

CLAUDE.md Grounding

Define project labels, “Definition of Done,” and CLI patterns in CLAUDE.md so agents understand the backlog context.

Agentic Triage Sweep

claude "Read the last 10 issues. Label by priority. Suggest 5-line fixes."

Spec-Driven Development

1. Raw ticket from FreeScout/Linear
2. copilot /specify → technical spec
3. copilot /plan → implementation steps
4. Claude Code executes, tests, submits PR

Recommended Hybrid Stack

• Inbound: FreeScout (best simple UI for non-technical users)
• Sync: Webhook mirrors “Developer Needed” tickets as Markdown files in GitHub repo
• Engine: Claude Code (CLI) + GitHub Copilot (IDE)
• Workflow: Open IDE → pull backlog → /ask for context → implement fix

Related

• custom-gsd-implementation — GAS framework for structured execution
• claude-code-hooks-workflow — Automated rule enforcement