Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write a clear Request for Proposal that gets comparable, high-quality vendor bids. Use when asked to write an RFP, a request for proposal/quote/tender, or to solicit and compare vendor proposals. Produces a complete RFP — background, scope of work, requirements, evaluation criteria with weights, submission instructions, and timeline — structured so responses are easy to compare apples-to-apples.
.claude/skills/mohitagw15856-rfp-writer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 138% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 79% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 92% | 0% |
A good RFP gets you proposals you can actually compare; a vague one gets a pile of incomparable sales decks. The trick is to specify the problem and the evaluation criteria clearly enough that vendors answer the same questions the same way. This skill writes that RFP — scoped, requirement-driven, and weighted — so selection is a defensible comparison, not a gut call.
Given "we need an RFP for a new CRM", produce the full RFP anyway — infer a sensible scope, requirements, and evaluation weights for that category, label assumptions, and bracket org-specifics (budget, dates, contacts) to fill in. Never withhold for missing detail.
Ask for these only if they aren't already provided (else infer and label):
| Criterion | Weight | What we're assessing | |---|---|---| | Capability / fit | 35% | meets mandatory + desirable requirements | | Price / TCO | 25% | total cost over the term, not just licence | | Support & SLAs | 15% | onboarding, support, uptime | | Security & compliance | 15% | data handling, certifications | | References / track record | 10% | proven delivery for similar orgs |
Procurement practice — requirement-driven scoping, weighted evaluation criteria set in advance, and structured questions for comparable bids.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 42,034 | 32,220 | -23% | 1 | 1 | 0% | 7,419 | 5,850 | -21% | 0 | 0 | — |
case-02 | fail→pass | 38,072 | 25,881 | -32% | 1 | 1 | 0% | 6,042 | 5,346 | -12% | 0 | 0 | — |
case-03 | fail→pass | 39,176 | 33,827 | -14% | 1 | 1 | 0% | 5,651 | 5,999 | +6% | 0 | 0 | — |
case-04 | pass→pass | 19,052 | 29,692 | +56% | 1 | 1 | 0% | 2,598 | 4,444 | +71% | 0 | 0 | — |
case-05 | pass→pass | 29,360 | 34,246 | +17% | 1 | 1 | 0% | 4,181 | 5,775 | +38% | 0 | 0 | — |
case-06 | pass→pass | 35,852 | 34,842 | -3% | 1 | 1 | 0% | 4,461 | 5,539 | +24% | 0 | 0 | — |
case-07 | fail→fail | 18,466 | 22,437 | +22% | 1 | 1 | 0% | 2,942 | 3,938 | +34% | 0 | 0 | — |
case-08 | fail→pass | 22,596 | 39,844 | +76% | 1 | 1 | 0% | 2,828 | 6,731 | +138% | 0 | 0 | — |
case-09 | pass→pass | 29,288 | 31,335 | +7% | 1 | 1 | 0% | 3,452 | 5,290 | +53% | 0 | 0 | — |
case-10 | pass→pass | 28,120 | 39,491 | +40% | 1 | 1 | 0% | 3,558 | 5,582 | +57% | 0 | 0 | — |
case-11 | fail→pass | 21,496 | 32,005 | +49% | 1 | 1 | 0% | 2,984 | 5,351 | +79% | 0 | 0 | — |
case-12 | pass→fail | 32,688 | 34,979 | +7% | 1 | 1 | 0% | 4,453 | 6,052 | +36% | 0 | 0 | — |
case-13 | fail→pass | 27,218 | 39,195 | +44% | 1 | 1 | 0% | 3,083 | 5,928 | +92% | 0 | 0 | — |
case-14 | fail→pass | 31,911 | 34,217 | +7% | 1 | 1 | 0% | 3,714 | 5,691 | +53% | 0 | 0 | — |
case-15 | fail→pass | 26,430 | 32,212 | +22% | 1 | 1 | 0% | 2,984 | 5,630 | +89% | 0 | 0 | — |
case-16 | fail→fail | 25,139 | 43,953 | +75% | 1 | 1 | 0% | 3,006 | 6,824 | +127% | 0 | 0 | — |
case-17 | fail→pass | 26,225 | 47,332 | +80% | 1 | 1 | 0% | 3,678 | 6,512 | +77% | 0 | 0 | — |
case-18 | fail→pass | 20,347 | 29,266 | +44% | 1 | 1 | 0% | 2,754 | 5,321 | +93% | 0 | 0 | — |
case-19 | fail→pass | 28,483 | 37,589 | +32% | 1 | 1 | 0% | 3,679 | 5,470 | +49% | 0 | 0 | — |
case-20 | fail→fail | 31,393 | 40,388 | +29% | 1 | 1 | 0% | 4,350 | 5,709 | +31% | 0 | 0 | — |
case-21 | fail→fail | 30,188 | 39,696 | +31% | 1 | 1 | 0% | 3,921 | 6,609 | +69% | 0 | 0 | — |
case-22 | pass→pass | 30,262 | 47,118 | +56% | 1 | 1 | 0% | 3,315 | 6,600 | +99% | 0 | 0 | — |
case-23 | pass→pass | 32,408 | 35,149 | +8% | 1 | 1 | 0% | 4,512 | 5,513 | +22% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +39 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.