lp-do-assessment-03-solution-selection — Solution Selection (ASSESSMENT-03)

Reads solution-space research results from the operator's deep research run, evaluates options against feasibility criteria, and produces a decision record with a shortlist of 1–2 viable options and elimination rationale for all others. An explicit decision record is required to continue to ASSESSMENT-04 (naming handoff).

Invocation

/lp-do-assessment-03-solution-selection --business <BIZ>

Required:

• --business <BIZ> — 3-4 character business identifier

Business resolution pre-flight: If --business is absent or the directory docs/business-os/strategy/<BIZ>/ does not exist, apply _shared/business-resolution.md before any other step.

Operating Mode

READ-THEN-DECIDE

This skill:

• Reads solution-space results and problem statement artifacts
• Builds an evaluation matrix across feasibility dimensions
• Produces a shortlist and decision record
• Applies a kill gate — if no viable option exists, stops the operator before ASSESSMENT-04

Required Inputs

Required:

• docs/business-os/strategy/<BIZ>/<YYYY-MM-DD>-solution-profile-results.user.md — operator-filled research output from /lp-do-assessment-02-solution-profiling

If no results file exists, stop: instruct the operator to complete the deep research run from /lp-do-assessment-02-solution-profiling first.

Also read if present:

• docs/business-os/strategy/<BIZ>/<YYYY-MM-DD>-problem-statement.user.md — use to verify options are still solving the stated problem

Steps

Step 1: Load and parse results

Read the results file. Extract all product-type options returned by the research tool. List them. If fewer than 3 options are present, flag as a thin research result and note that the shortlist may be unreliable.

Step 2: Build the evaluation matrix

For each option, score across these four criteria:

1. Launch feasibility — can a small team launch this within 6 months with realistic capital?
2. Distribution path — is there a known, accessible channel to reach the target user group?
3. Regulatory/compliance burden — are there licensing, certification, or compliance requirements that would materially delay or block launch?
4. Problem-solution fit — does this option directly address the stated problem in <YYYY-MM-DD>-problem-statement.user.md, or does it require the problem to be reframed?

Score each: Pass / Flag / Fail. A single Fail does not automatically eliminate an option — use judgment based on all four together. Document the rationale for each score; do not leave scores unexplained.

Step 3: Apply elimination logic

Eliminate options that:

• Score Fail on Launch feasibility AND Fail on Distribution path
• Score Fail on Regulatory burden (hard legal block, not just complexity)
• Score Fail on Problem-solution fit (option solves a different problem)

For each eliminated option, write a 1-sentence elimination rationale. "Not interesting" is not a valid rationale — cite the specific criterion that failed.

Step 4: Select shortlist

From remaining options, select 1–2 to carry forward to ASSESSMENT-04. If two options remain and both are genuinely viable, include both — the naming research phase can handle parallel tracks. Do not reduce to one arbitrarily.

Apply the problem-fit check: verify that each shortlisted option still addresses the problem statement as written. If an option requires reframing the problem, note this explicitly in the decision record.

Step 5: Apply kill gate

If no options survive elimination, apply the kill gate:

Explicit decision record required to continue to ASSESSMENT-04. No viable option identified from current solution-space results. Do not advance to ASSESSMENT-04 (naming handoff). Options:

1. Re-run /lp-do-assessment-02-solution-profiling --business <BIZ> with a refined problem scope or broader option set
2. Return to /lp-do-assessment-01-problem-statement --business <BIZ> to reassess the problem statement

Write the kill gate output to the artifact and stop. Do not proceed to ASSESSMENT-04.

Step 6: Write and save decision record

Assemble all outputs into the decision record format below. Save to the output path.

Output Contract

Path: docs/business-os/strategy/<BIZ>/<YYYY-MM-DD>-solution-decision.user.md

Format:

# Option Selection Decision — <BIZ>

## Evaluation Matrix

| Option | Launch Feasibility | Distribution Path | Regulatory Burden | Problem Fit | Status |
|---|---|---|---|---|---|
| <option name> | Pass/Flag/Fail | Pass/Flag/Fail | Pass/Flag/Fail | Pass/Flag/Fail | Shortlisted/Eliminated |

## Elimination Log

- **<option>**: [1-sentence rationale citing specific criterion]
- **<option>**: [1-sentence rationale citing specific criterion]

## Shortlist

1. **<option name>** — [2-3 sentence rationale for selection; any caveats]
2. **<option name>** (if applicable) — [2-3 sentence rationale]

## Decision

**Status:** GO / NO-GO

**Carry forward:** <option name(s)>

[If GO: "Proceed to `/lp-do-assessment-04-candidate-names --business <BIZ>` with the option(s) above."]
[If NO-GO: Kill gate language — see Step 5.]

Downstream consumer: /lp-do-assessment-04-candidate-names --business <BIZ> (ASSESSMENT-04) reads this file to scope the naming research to the selected option(s).

Quality Gate

Before saving, verify:

• All options from results file appear in the evaluation matrix
• Every score has a documented rationale — no unexplained Pass/Flag/Fail
• Elimination log contains one entry per eliminated option with specific criterion cited
• Shortlist contains 1–2 options (not zero unless kill gate applies, not more than 2)
• Kill gate applied if no viable options remain
• ## Decision section contains explicit GO/NO-GO verdict
• Shortlisted options align with the problem statement from <YYYY-MM-DD>-problem-statement.user.md
• Artifact saved to correct path before completion message

Red Flags

Invalid outputs — do not emit:

• Evaluation matrix with unexplained scores ("it seemed weak")
• Elimination rationale of "not interesting" or "too crowded" without citing specific criteria
• Shortlist of zero options without triggering the kill gate
• Decision section missing or containing only "TBD"
• Kill gate suppressed — if no viable option exists, the gate MUST fire; do not advance the operator to ASSESSMENT-04

Integration

Upstream (ASSESSMENT-02): Reads docs/business-os/strategy/<BIZ>/<YYYY-MM-DD>-solution-profile-results.user.md from the operator's deep research run via /lp-do-assessment-02-solution-profiling.

Downstream (ASSESSMENT-04): /lp-do-assessment-04-candidate-names --business <BIZ> reads <YYYY-MM-DD>-solution-decision.user.md to scope naming research to the shortlisted option(s). The naming skill should be run with the shortlisted option names provided as context.