What problem does it solve? Keeping issue trackers accurate and well-structured is tedious and error-prone: issues get created without clear verification criteria, dependency relationships drift, and bulk edits happen without review. This Skill shapes work into verifiable issues, maintains native parent/sub-issue/blocker graphs in GitHub or Linear, and gates every tracker mutation behind a complete preview and explicit approval. ## Core Features & Use Cases - Issue shaping and decomposition: Draft issues from a Problem/Scope/Verification template and split large work only into independently deliverable, verifiable leaves. - Dual-provider support: Operates against GitHub (via authenticated gh CLI) or Linear (via connected MCP tools or the Orca CLI), with a validated per-repository config at .agents/managing-issues.json. - Graph, readiness, and completion management: Traverse issue families, derive readiness (needs-discovery, needs-planning, ready), compute the Ready Frontier, and verify completion against evidence before closing. - Safe mutation lifecycle: Every batch is previewed with exact effects, directly approved, revalidated before each write, applied once, and read back with applied/failed/indeterminate classification. - Use Case: Ask the agent to break a feature request into GitHub sub-issues with blocker relationships; it drafts each issue, previews the full ordered batch of creates and edges, and applies them only after your approval. ## Quick Start Use the managing-issues skill to draft and create a GitHub issue for the login timeout bug with clear verification criteria, then show me the preview before writing anything.
