tp-spike-plan

Spike Experiment Plan

Turn a spike design into a structured experiment plan with hypothesis-driven tasks.

Arguments:

• {spike-name} (required) — must match an existing directory under three-pillars-docs/tp-designs/.
• --auto (optional) — autonomous mode. Skips user confirmation, self-reviews plan, logs decisions to decisions.md. See skills/_shared/auto-mode.md for convention.

Prerequisites

• three-pillars-docs/tp-designs/{spike-name}/design.md must exist and be spike-flavored (exploratory, not implementation). If not, tell the user to create one first and stop.

Steps

1. Run first-run preflight per skills/_shared/first-run.md.

2. Run collaboration preflight per skills/_shared/collaboration.md with phase: "spike-plan". This verifies the branch and refreshes the lock for this spike. Honor --force-takeover if passed. In --auto mode, do not prompt — if the lock is held by another developer, log the conflict to decisions.md and stop.

3. Read design.md from the design directory.

4. Identify the experiments described in the design. Group them into phases — each phase explores one facet of the spike.

5. Break each phase into 2-5 tasks. Each task is a single experiment:

   • Hypothesis-driven — states what you expect before trying
   • Has a clear evaluation method — how to judge the result
   • Produces a Deliverable at the phase level for user review

6. Write three-pillars-docs/tp-designs/{spike-name}/plan.md using this exact format:

# <Spike Name> — Experiment Plan

## Phase 1: {description}
**Deliverable**: What the user reviews at phase end

### Task 1.1: {description}
**Hypothesis**: What we expect to happen
**Try**: What to implement/run
**Evaluate**: How to judge the result (visual check, metric, test)
**Status**: Pending

### Task 1.2: ...

## Phase 2: {description}
**Deliverable**: ...

1. Confirm the plan:
   • Normal mode: Walk through it with the user. Adjust if they want to reorder, add, or drop tasks.
   • --auto mode: Self-review the plan against design.md — verify every experiment in the design maps to at least one task, success criteria are testable, and phases are logically ordered. Log a decision entry to decisions.md summarizing plan structure choices (how experiments were grouped into phases, any design experiments that were split or combined, and why). Auto-approve and continue.
2. Commit the artifact per skills/_shared/commit-after-work.md. Artifact paths to stage:
   • three-pillars-docs/tp-designs/{spike-name}/plan.md
   • three-pillars-docs/tp-designs/{spike-name}/lock.json (if refreshed) Commit message: Spike: {spike-name} plan.
3. Tell the user (or log, in auto mode) the next step is to begin executing the experiment phases.