agentskills.codes
ST

start-next-task

Use when you want to pick the next Linear issue to work on. Fetches unstarted todo issues from Linear, displays them, lets you select one, and auto-updates your local task tracking. Provides implementation guidance.

Install

mkdir -p .claude/skills/start-next-task && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14914" && unzip -o skill.zip -d .claude/skills/start-next-task && rm skill.zip

Installs to .claude/skills/start-next-task

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use when you want to pick the next Linear issue to work on. Fetches unstarted todo issues from Linear, displays them, lets you select one, and auto-updates your local task tracking. Provides implementation guidance.
215 chars✓ has a “when” trigger

About this skill

Start Next Task Skill

Overview

This skill streamlines your issue selection and task startup workflow by:

  1. Fetching unstarted/todo issues from Linear
  2. Displaying them with key details (ID, title, status)
  3. Updating tasks.md with your selected issue
  4. Reviewing requirements.md and plan.md for project-level specs
  5. Providing implementation guidance based on issue description

Workflow Steps

Step 1: Fetch Unstarted Issues

Use Linear MCP to list all issues with status "Todo" filtered by the current team.

Pseudo-code:

issues = linear.list_issues(
  status: ["Todo", "In Progress"],
  orderBy: "priority",
  limit: 10
)

Step 2: Present Selection

Display issues in a numbered list with:

  • Issue ID (e.g., LIST-5)
  • Title
  • Description preview (first 100 chars)
  • Status

Example Output:

Available issues:

1. [LIST-5] Update list item UI component
   Status: Todo
   Improve visual hierarchy and spacing

2. [LIST-6] Add drag-and-drop reordering
   Status: Todo
   Implement reordering for items within a list

Step 3: User Selection

Ask user to pick an issue number (1, 2, etc.).

Step 4: Fetch Full Details

Once selected, fetch complete issue details:

  • Full description
  • Acceptance criteria (if in description)
  • Dependencies
  • Related tasks

Step 5: Update tasks.md

Parse the selected issue and add/update entry in tasks.md:

  • Add issue ID (e.g., [LIST-5]) to the section
  • Mark relevant checkboxes as current work
  • Link to Linear issue if not already present

Step 6: Implementation Guidance

Extract implementation hints from issue description:

  • List acceptance criteria
  • Highlight dependencies or prerequisites
  • Point to relevant source files in the codebase

How to Use

In VS Code chat, type:

/select-next-issue

Or ask:

Pick the next issue I should work on from Linear.

Integration Points

Linear MCP

  • Requires Linear workspace access
  • Fetches issues by status and team
  • Retrieves full issue details

Local Task Tracking

  • Reads from tasks.md
  • Updates section headers with issue ID
  • Marks related subtasks

Git

  • Creates feature branches following pattern: {ISSUE_ID}-{kebab-case-title}
  • Provides branch creation guidance

Example Session

User: /select-next-issue

Skill: Fetching your unstarted Linear issues...

[Lists 5 issues]

User: I'll pick #2

Skill: Selected: LIST-6 Add drag-and-drop reordering

[Updates tasks.md locally]
[Suggests git branch: LIST-6-add-drag-and-drop]

Skill: Ready to start! Here's what needs implementing:

Acceptance Criteria:
- [ ] Drag items within a list to reorder
- [ ] Position persists to database
- [ ] Animation smooth on both iOS and Android

Prerequisites:
- ✓ SQLDelight database layer (LIST-2 complete)
- ✓ Jetpack Compose layout basics

Next: Create a git branch and start coding.

Assets & Tools

Linear MCP Integration

This skill uses the Linear MCP server to:

  • Query issues by status
  • Retrieve issue details
  • Access issue descriptions and metadata

No External Scripts Required

All logic is handled by the AI agent with Linear MCP integration.

Search skills

Search the agent skills registry