Agent-Skills.md

Agent Skills: Linear Triage

Manage Linear projects — triage boards, create issues and epics, enforce labels and metadata, audit dependencies. Use when user wants to triage issues, create/review tickets, clean up a board, manage labels/priorities/milestones, or write epics with sub-issue decomposition.

UncategorizedID: enitrat/skill-issue/linear-triage

Repository

enitrat/skill-issue
enitrat
142

Install this agent skill to your local

pnpm dlx add-skill https://github.com/enitrat/skill-issue/tree/HEAD/skills/linear-triage

Skill Files

Browse the full folder contents for linear-triage.

Download Skill

Loading file tree…

skills/linear-triage/SKILL.md

Skill Metadata

Name
linear-triage
Description
Manage Linear projects — triage boards, create issues and epics, enforce labels and metadata, audit dependencies. Use when user wants to triage issues, create/review tickets, clean up a board, manage labels/priorities/milestones, or write epics with sub-issue decomposition.

Linear Triage

Manage and triage issues in Linear using mcp__claude_ai_Linear__* MCP tools. All mutations follow prepare → confirm → execute: present proposed changes, wait for approval, then apply.

Reference docs

  • TRIAGE.md — board-level triage, health dashboard, batch operations
  • ISSUE-WRITING.md — Why/What/How template, quality scoring, anti-patterns
  • EPIC-WRITING.md — epic structure, decomposition, sub-issue boundaries
  • LABELS.md — label taxonomy and enforcement rules

Core principles

  1. Every issue earns its place. No Why, no concrete What, no testable acceptance criteria → flag it.
  2. Structure over volume. A 2000-word wall of text is noise. Why/What/How with clear sections beats prose dumps.
  3. Metadata is the filtering API. Missing labels, priorities, or milestones make the board unqueryable.
  4. Lock the contract, free the implementation. Issues define outcomes and boundaries. The implementer decides internal code structure.
  5. Prepare, confirm, execute. Never mutate without showing the user exactly what will change.

Invocation

The user invokes /linear-triage then describes what they want.

Triage requests → follow TRIAGE.md:

  • "Triage the SDK project" / "What needs attention?"
  • "Review SDK-42" / "Flag issues missing labels"
  • "Check dependencies on the v0.12 milestone"

Issue creation → follow ISSUE-WRITING.md:

  • "Create a ticket for X" / "Write up an issue for the auth bug"

Epic creation → follow EPIC-WRITING.md:

  • "Create an epic for the credential refactor"
  • "Decompose this work into sub-issues"

Label/metadata questions → follow LABELS.md:

  • "What label should this get?" / "Fix all missing labels"

Workflow: Create issue

  1. Clarify with the user: what problem, what outcome, what's out of scope.
  2. Draft the issue using the Why/What/How template from ISSUE-WRITING.md.
  3. Score against the quality checklist (must score ≥ 7 for ready).
  4. Apply metadata: priority, category label, milestone, assignee.
  5. Present the full issue for confirmation before creating via save_issue.

Workflow: Create epic

  1. Clarify: what problem, what "done" looks like, what's NOT in scope.
  2. Write the epic using the 6-section structure from EPIC-WRITING.md.
  3. Decompose into sub-issues (vertical slices, risk-ordered, 1–5 days each).
  4. Write each sub-issue using Why/What/How from ISSUE-WRITING.md.
  5. Score epic ≥ 7. Apply epic label + category label, priority, milestone.
  6. Present epic + sub-issues for confirmation before creating.

Workflow: Triage

Follow TRIAGE.md for board-level or single-issue triage.

Comments

When posting triage comments to Linear:

**Triage note** (generated by AI)

[content]