agentleFS
devoxxdevoxx/DevoxxGenieIDEAPlugin/.claude/skills/start-task/SKILL.md

start-task

Skill689 starsChanged 8 days agoMITView on GitHub

Create a feature branch for a backlog task, switch to it, and start implementation. Use when the user says "start task-123", "work on task-123", "implement task-123", or invokes /start-task with a task ID.

16 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 10 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create a feature branch for a backlog task and begin implementation.

    line 17
  2. Very specific:

    Use mcp__backlog__task_view to read the full task details: title, description, acceptance criteria, and any referenced files.

    line 24
  3. Very specific:

    If the task is not found, use mcp__backlog__task_search to locate it.

    line 25
  4. Distinctive:

    Check git status. If there are uncommitted changes, warn the user and ask whether to stash them before proceeding.

    line 30
  5. Distinctive:

    Ensure we are on the main branch.

    line 31
  6. Very specific:

    Create and switch to the branch: git checkout -b <branch-name>.

    line 38
  7. Very specific:

    Use mcp__backlog__task_edit to set the task status to In Progress.

    line 42
  8. Distinctive:

    Read all files referenced in the task, plus any related code you need to understand.

    line 47
  9. Distinctive:

    Create a plan and present it to the user for approval before writing code.

    line 48
  10. Distinctive:

    After implementation, run the full verification suite:

    line 50
  11. Distinctive:

    Always create the branch from an up-to-date main branch.

    line 63
  12. Distinctive:

    Never start implementation without showing the plan to the user first.

    line 64
  13. Distinctive:

    Follow existing code patterns and conventions in the project.

    line 65
  14. Very specific:

    Do not use interactive git flags (-i).

    line 66
  15. Distinctive:

    If the working tree is dirty, never silently discard changes.

    line 67

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Start Task

Context

  • Current branch: !git branch --show-current
  • Working tree status: !git status --short
  • Existing branches: !git branch --list 'feature/task-*'

Your task

Create a feature branch for a backlog task and begin implementation.

The user will provide a task ID (e.g. task-123 or 123). If no task ID is provided, check the argument $ARGUMENTS for the task reference.

Step 1 — Load the task

  • Normalize the input: if the user gave just a number like 123, treat it as task-123.
  • Use mcp__backlog__task_view to read the full task details: title, description, acceptance criteria, and any referenced files.
  • If the task is not found, use mcp__backlog__task_search to locate it.
  • Print a brief summary of the task for the user.

Step 2 — Ensure a clean working tree

  • Check git status. If there are uncommitted changes, warn the user and ask whether to stash them before proceeding.
  • Ensure we are on the main branch. If not, ask the user if they want to switch.

Step 3 — Create and switch to a feature branch

  • Pull latest changes on main: git pull --rebase.
  • Derive a branch name from the task: feature/task-{id}-{slugified-title} (lowercase, hyphens, max ~60 chars).
    • Example: task 113 "Split app.css into native CSS modules" → feature/task-113-split-css-modules
  • Create and switch to the branch: git checkout -b <branch-name>.

Step 4 — Mark the task as in-progress

  • Use mcp__backlog__task_edit to set the task status to In Progress.

Step 5 — Plan and implement

  • Analyze the task description and acceptance criteria carefully.
  • Read all files referenced in the task, plus any related code you need to understand.
  • Create a plan and present it to the user for approval before writing code.
  • Once approved, implement the changes following the project's patterns and conventions.
  • After implementation, run the full verification suite:
    npm run lint && npm run typecheck && npm test && npm run build:app && npx playwright test
  • Fix any issues found by the verification suite.

Step 6 — Report

  • Summarize what was implemented and which acceptance criteria were addressed.
  • Remind the user they can use /close-task-commit-push-pr when ready to ship.

Rules

  • Always create the branch from an up-to-date main branch.
  • Never start implementation without showing the plan to the user first.
  • Follow existing code patterns and conventions in the project.
  • Do not use interactive git flags (-i).
  • If the working tree is dirty, never silently discard changes.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.