Write comprehensive task breakdowns assuming the expert who is going to implement the specs has zero context for our project and questionable taste. Document everything they need to know: which existing files to check, which files to touch for each task and what changes to make to them. Give them the whole plan as bite-sized tasks. DRY. YAGNI. TDD.
Assume they are a skilled worker, but know almost nothing about our toolset or problem domain. Assume they don't know how to verify they are doing the right thing.
Analyze available skills and propose creating new skills if needed. If you propose creating new skills, you MUST create them before creating the task breakdown.
Announce at start: "I'm using the task-breakdown skill to create a plan."
Presenting the tasks:
Once you believe you have the full task breakdown, present the tasks one-by-one to the user
Ask after each task whether it looks right so far
Be ready to go back and clarify if something doesn't make sense
Save tasks to:docs/YYYY-MM-DD-<feature-name>-tasks.md
Bite-Sized Task Granularity
Each step is one action (2-5 minutes):
"Write the failing test" - step
"Run it to make sure it fails" - step
"Implement the minimal code to make the test pass" - step
Run: pytest tests/path/test.py::test_name -v Expected: FAIL with "function not defined"
Step 3: Write minimal implementation
def function(input):
return expected
Step 4: Cleanup code changes Use skill(s) if available to cleanup code changes
Step 5: Review code changes Use skill(s) if available to review code changes. Make sure code follows the project's coding standards and aligns with the specs and the task breakdown.
