Глубокий разбор GitHub Issues с анализом архитектуры

★ 7.4 · documentation

Issue Analysis is a Claude Code skill that performs deep analysis of a single GitHub issue — exploring the codebase, assessing architectural implications, planning implementation approaches, and posting a structured comment directly on the issue with recommended labels. Reach for it when a bug report or feature request feels too complex to start on; when you need to estimate implementation effort and spot hidden architectural risks before committing; when prioritizing a backlog and you need a clear picture of how open issues relate to each other. Built on Bash, Read, Glob, Grep, WebFetch, and WebSearch tooling with the Opus model, the skill follows a strict pipeline: fetch issue data via GitHub CLI, grep and read src/ for related code, verify technical claims against the actual source, web-search external APIs and library versions for accuracy, scan up to 50 open issues for priority context, assign a classification label (needs-choices / ready-to-implement / needs-info) plus a priority level, draft the full comment for user review — and only post after explicit approval. The flow is analysis and planning only, not implementation; writing code is explicitly out of scope. The skill is wired to the homeassistant-ai/ha-mcp repository and requires a configured GitHub CLI authenticated with comment and label permissions.