Test writing-plans skill compliance with a Sonnet subagent
How this command is triggered — by the user, by Claude, or both
Slash command
/superpowers-extended-cc:test-writing-plansThe summary Claude sees in its command listing — used to decide when to auto-load this command
Test the writing-plans skill by spawning a Sonnet subagent that executes it with a dummy task. Verify the subagent follows the mandatory workflow. ## Execution 1. Determine the absolute path of the current project root (where this command lives) 2. Spawn a Sonnet subagent with this prompt (replace `PROJECT_ROOT` with the actual path): ## After Subagent Completes 1. Check the subagent's compliance report (the 5 yes/no answers) 2. Verify files were created in the right location: 3. Clean up test artifacts: 4. Delete any native tasks created by the test subagent ## Expected Resu...
Test the writing-plans skill by spawning a Sonnet subagent that executes it with a dummy task. Verify the subagent follows the mandatory workflow.
PROJECT_ROOT with the actual path):Task(
subagent_type="general-purpose",
model="sonnet",
prompt="""You are testing a skill. Read the skill file and follow it EXACTLY as written. This is a test run.
First, read the skill at: PROJECT_ROOT/skills/writing-plans/SKILL.md
Then execute it with this task: "Add a hello-world CLI command to the superpowers project that prints 'Hello from superpowers'"
The project is at PROJECT_ROOT
Follow every step in the skill document in exact order. Do NOT skip any steps.
IMPORTANT: At the end, report which tools you called in order, specifically:
1. Did you call TaskList as the first tool? (yes/no)
2. Did you call EnterPlanMode or ExitPlanMode at any point? (yes/no)
3. Did you call AskUserQuestion at the end? (yes/no)
4. Did you save the plan to docs/plans/YYYY-MM-DD-*.md? (yes/no)
5. Did you save .tasks.json next to the plan? (yes/no)"""
)
ls docs/plans/*hello*
rm docs/plans/*hello*
All 5 must pass for the test to succeed:
npx claudepluginhub rishikanthc/ml-superpowers --plugin superpowers-extended-cc/test-skillTests Claude Code skills through RED/GREEN/REFACTOR TDD phases in fresh subagents for validation, checkpoint enforcement, and priming bias prevention. Supports [skill-path] [--phase red|green|refactor].
/testGenerates and runs trigger tests, functional tests, and performance baselines for a skill given path to folder or SKILL.md.
/plan-structure-templateGenerates structured feature implementation plan as multi-file folder: _index.md overview with YAML tasks plus one detailed task file each.
/test-frameworkRuns four-layer test framework for Claude Code plugin skills: validates structure, tests trigger accuracy, runs session scenarios, and compares skill value.