starting-the-task

lambdamechanic's avatarfrom lambdamechanic

A short checklist for kicking off work effectively: plan, branch, track with bd, and set up validation.

⭐0stars🔀1forks📁View on GitHub🕐Updated Dec 5, 2025
[SOPs and Playbooks]

When & Why to Use This Skill

This Claude skill provides a standardized developer workflow checklist designed to streamline the transition from planning to execution. It ensures that every technical task is grounded, observable, and reversible by enforcing best practices such as environment validation, automated branch management, and early PR creation. By following this rigorous SOP, teams can maintain high code quality and ensure project tracking is always up to date.

Use Cases

  • Standardizing Task Initialization: Automatically following a consistent set of steps (planning, branching, and status updates) whenever a new development ticket is started.
  • Environment Health Checks: Preventing work on broken baselines by verifying project test suites (e.g., cargo test) before diverging from the main branch.
  • Enhanced Project Visibility: Improving stakeholder communication by immediately creating draft PRs and updating issue trackers (bd) to reflect 'in-progress' status.
  • Risk Mitigation: Ensuring all changes are proofed through early test planning and Git best practices to allow for easy reverts and clear audit trails.
namestarting-the-task
description"A short checklist for kicking off work effectively: plan, branch, track with bd, and set up validation."

Starting the Task

Use this checklist any time you begin work on a bd issue (or whenever the user says “start working on …”). It ensures the effort is grounded, observable, and reversible.

  1. Understand the Task

    • Read the bd issue (and any linked discoveries or docs) end-to-end.
    • Confirm acceptance criteria, dependencies, and blockers. Ask clarifying questions before touching code.

  2. Environment Readiness

    • Run cargo test (or the project’s canonical smoke suite) on main to ensure the baseline is green before you diverge.
    • If tests fail, stop and coordinate—don’t start stacking new work on a broken branch.

  3. Branch & Tracking

    • Create a fresh branch for the effort, even if you expect a small change.
      git checkout -b <short-task-name>
    • bd update <id> --status in_progress --notes "Starting work" so the tracker reflects the new ownership/status.
    • If multiple repos are involved, repeat for each.

  4. Open a Draft PR (immediately)

    • As soon as you create the branch, open a draft PR so CI runs and stakeholders can track progress. If a PR already exists for the parent effort, continue on that one.
    • Source the PR title/description from the bd issue: include its summary, acceptance criteria, and a link/reference to the bd id.
    • Preferred command (uses GitHub CLI):
      scripts/pr-draft.sh "<short title incl. bd id>" [body.md]
      Fallback: gh pr create --draft --title "..." --body-file body.md after git push -u origin <branch>.

  5. Sync & Tooling

    • Ensure git pull --rebase (or equivalent) so you’re working from the latest remote main.
    • Verify gh auth status so PR creation won’t fail later.

  6. Plan Tests Early

    • Decide how you’ll prove the change works (unit tests, integration, manual steps). Capture this in notes or the eventual PR description so reviewers know what to expect.

Only after these steps are complete should you begin coding. If circumstances prevent any step (e.g., CI down, baseline broken), note it explicitly in bd and get confirmation before proceeding.