Idea scorecard
Skill eltntawy/idea-discovery-playbook/skills/idea-scorecard
16 agent skills to validate software ideas before you build — Mom Test interviews, JTBD, market research, pretotype tests, unit economics, and go/no-go scorecards. For Cursor, Claude Code & skills.sh.
npx -y skills add eltntawy/idea-discovery-playbook --skill idea-scorecardAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Scores software ideas across validation stages using an evidence hierarchy, assumption ledger, weighted go, conditional go, pivot, or kill thresholds. Use when the user asks to score an idea, make a go/no-go decision, or validate before build.
SKILL.md
3.5 KB, as published. Nobody here has run it
Idea Scorecard
Evidence hierarchy (canonical)
Weakest to strongest — map all memos and experiments to these levels:
| Level | Label | Examples |
|---|---|---|
| 1 | Awareness | Views, impressions |
| 2 | Interest | Opinions, compliments, hypothetical praise |
| 3 | Intent | Signups, meetings, waitlists with context |
| 4 | Behavior | Repeat use, workflow adoption, design-partner time |
| 5 | Revenue | Payment, deposits, LOIs, pre-orders at target price |
behavior-led-validation uses L0–L5 labels that map to this hierarchy (L0–L1 ≈ 1–2, L2 ≈ 3, L3 ≈ 4, L4–L5 ≈ 5).
Assumption ledger
Before scoring, list open claims from assumption-map or prior memos:
| Assumption | Evidence tier (1–5) | Test | Status |
|---|---|---|---|
| open / validated / killed |
Cap Revenue and Channel stage scores when critical assumptions remain open with tier ≤ 2.
Weighted scorecard
| Stage | Score 1-10 | Weight | Weighted | Evidence summary |
|---|---|---|---|---|
| Problem | 1.5 | |||
| Solution | 1.5 | |||
| Market | 1.0 | |||
| Channel | 1.0 | |||
| Revenue | 1.5 | |||
| Founder–market fit | 0.5 | Domain, access, distribution, constraints | ||
| Scale | 0.5 | |||
| Total | /75 |
Total max: sum of (score × weight). Weights sum to 7.5; perfect 10s → 75.
Decision thresholds
- 60–75: Strong go
- 45–59: Conditional go — close named gaps first
- 30–44: Pivot
- 0–29: Kill or restart
Recalculate thresholds if you omit Founder–market fit (max 70, bands: 56 / 42 / 28 / 0).
Kill cues
- Problem confirmed by fewer than half of relevant interviews
- No paying customers after four weeks of revenue testing
- LTV:CAC below 2:1 in the base case (see
revenue-modelbands) - All acquisition channels exceed target CAC
feasibility-gaterecords Reject or Wait
LTV:CAC 2:1 to < 3:1 does not auto-kill; use Conditional go and document the economics gap.
Output
Save to discovery/<topic-slug>/scorecards/YYYY-MM-DD-<topic-slug>-scorecard.md using TEMPLATE.md.
Do not recommend implementation unless the decision is Strong go or Conditional go with explicit remaining gaps.
After scoring, offer discovery-synthesis to compile a researcher pack.