iterate - SKILL.md Agent Skill

name: iterate description: Run the AI-SHIPR Iteration Planner. Close one cycle and open the next. Run after a measurement window completes or hypothesis is validated/invalidated.

You are running the AI-SHIPR Iteration Planner for a product manager.

Read Settings.md first.

If product_mode: multi: Ask which product this iteration covers before reading any files. Use [product-name]/ as the path prefix for product-specific files below.

Read these files before proceeding:

All files in P-Proof/ (or [product-name]/P-Proof/ in multi mode)
Learning.md
All files in H-Hypotheses/ (or [product-name]/H-Hypotheses/ in multi mode)
All files in I-Initiatives/ (or [product-name]/I-Initiatives/ in multi mode)
S-Strategy/Strategic-Bets.md (or [product-name]/S-Strategy/Strategic-Bets.md in multi mode)

Ask for the initiative name if not provided. Then generate the Iteration Plan below.

Iteration Plan

Initiative: [Name] Cycle: [Sprint / Quarter / Date range] Hypothesis status: Validated / Partially Validated / Invalidated / Inconclusive

Cycle Close Summary

What Shipped:

[Feature / change 1]
[Feature / change 2]

What the Data Said:

[Honest summary of Performance-Tracker findings] Confidence: [High / Medium / Low]

What We Learned:

[What this cycle taught us about the user, the problem, or the product] Source: Learning.md entry from [date]

Next Cycle Decision

Based on proof and learning, assess these directions:

Option	Description	Basis	Recommended?
Double down	Expand on what was validated	[Evidence]	Yes / No
Pivot	Change approach — same problem, different solution	[What the data revealed]	Yes / No
Iterate	Refine the shipped solution	[What partially worked]	Yes / No
Kill	Stop investing — hypothesis invalidated	[Evidence]	Yes / No
Explore	Validated result opened a new question	[Signal]	Yes / No

Recommended Direction: [Double Down / Pivot / Iterate / Kill / Explore]

Rationale:

[Why this direction — tied to the data, the learning, and the strategic bet]

What changes:

[What is different in the next cycle]
[What we are dropping]
[What we are adding]

Next Cycle Initiative Candidates

Initiative	Source	Strategic Bet	Priority Signal	Hypothesis Ready?
[Name]	Backlog / New signal / Partial from this cycle	[Bet #]	High / Med / Low	Yes / No

Next Hypothesis Candidate

If the recommended direction requires a new hypothesis:

We believe [target user] will [behavior] because [reason]. We will measure [metric] over [time window]. Success: [threshold]. Failure: [threshold].

→ Run Hypothesis-Builder to formalize this into an HYP file.

Learning.md Entry to Write

[Date] — Iteration close: [Initiative name] Outcome: [Validated / Invalidated / Partial] Key learning: [What this cycle taught us] Next direction: [Double down / Pivot / Iterate / Kill / Explore] Next hypothesis: [One sentence — or "TBD — discovery needed"]

Actions Required

Update initiative Stage to: Iterating or Defined
Update H-Hypotheses file status: Validated / Invalidated / Archived
Create new initiative file if direction is "Double Down" or "Pivot"
Run Hypothesis-Builder if next cycle requires a new testable hypothesis
Run Sprint-Planner to load next cycle into sprint structure
Add learning entry to Learning.md