Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write a tight Statement of Work (SOW) that prevents scope creep and payment disputes. Use when asked to write a SOW, a scope of work, a project agreement, or to formalise what was agreed after a proposal. Produces an SOW — scope (and explicit exclusions), deliverables with acceptance criteria, timeline & milestones, payment schedule, assumptions, change-control, and terms. The contract layer after the proposal sells.
.claude/skills/mohitagw15856-statement-of-work/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 81% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 60% | 0% |
The proposal wins the deal; the SOW protects it. Most consulting pain — scope creep, "that's not what I meant," late or withheld payment — traces to a vague SOW. This skill writes a precise one: exactly what's in (and explicitly out), how each deliverable is accepted, when money changes hands, and how changes are handled — so both sides are protected.
Ask for these only if they aren't already provided:
consulting-proposal).Between: provider] and client] · Effective: date]
1. Scope — what will be done, specifically. Then explicit exclusions ("Out of scope: …") — the most valuable section; unsaid scope is assumed-included by clients.
2. Deliverables & acceptance criteria — each deliverable with how it's accepted (the objective bar, and a review window — e.g. "approved, or feedback within 5 business days, else deemed accepted").
| Deliverable | Acceptance criteria | Due | |---|---|---|
3. Timeline & milestones — phases, dates, and client dependencies (their inputs/approvals — and what happens to the timeline if they slip).
4. Payment schedule — amounts tied to milestones/dates, invoicing terms, and late-payment terms. Deposit up front where appropriate.
5. Assumptions — what the plan and price depend on (access, environments, responsiveness) — so a broken assumption is a change, not a fight.
6. Change control — how scope changes are requested, priced (the change rate), and approved in writing before work proceeds. This is the anti-scope-creep clause.
7. Terms — IP/ownership (on payment), confidentiality, termination, liability — flag that legal should review for material engagements.
Statement-of-work / contracting practice — explicit scope + exclusions, acceptance criteria, milestone payments, change control.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 25,506 | 29,656 | +16% | 1 | 1 | 0% | 3,950 | 5,192 | +31% | 0 | 0 | — |
case-02 | fail→pass | 31,950 | 35,423 | +11% | 1 | 1 | 0% | 4,050 | 5,216 | +29% | 0 | 0 | — |
case-03 | fail→pass | 29,875 | 24,027 | -20% | 1 | 1 | 0% | 3,566 | 4,274 | +20% | 0 | 0 | — |
case-04 | pass→pass | 11,517 | 15,001 | +30% | 1 | 1 | 0% | 2,360 | 2,643 | +12% | 0 | 0 | — |
case-05 | pass→pass | 19,961 | 23,940 | +20% | 1 | 1 | 0% | 3,404 | 3,931 | +15% | 0 | 0 | — |
case-06 | pass→pass | 18,249 | 23,633 | +30% | 1 | 1 | 0% | 2,468 | 4,061 | +65% | 0 | 0 | — |
case-07 | fail→fail | 6,671 | 23,848 | +257% | 1 | 1 | 0% | 1,214 | 4,255 | +250% | 0 | 0 | — |
case-08 | fail→fail | 22,934 | 24,722 | +8% | 1 | 1 | 0% | 2,741 | 4,129 | +51% | 0 | 0 | — |
case-09 | fail→pass | 17,607 | 24,762 | +41% | 1 | 1 | 0% | 2,363 | 4,285 | +81% | 0 | 0 | — |
case-10 | fail→pass | 22,021 | 23,069 | +5% | 1 | 1 | 0% | 2,465 | 3,935 | +60% | 0 | 0 | — |
case-11 | fail→pass | 19,093 | 19,183 | +0% | 1 | 1 | 0% | 2,532 | 4,406 | +74% | 0 | 0 | — |
case-12 | fail→fail | 20,367 | 20,804 | +2% | 1 | 1 | 0% | 2,063 | 3,478 | +69% | 0 | 0 | — |
case-13 | fail→fail | 19,552 | 28,995 | +48% | 1 | 1 | 0% | 2,508 | 4,000 | +59% | 0 | 0 | — |
case-14 | fail→pass | 13,608 | 25,516 | +88% | 1 | 1 | 0% | 1,403 | 4,534 | +223% | 0 | 0 | — |
case-15 | fail→fail | 23,881 | 22,276 | -7% | 1 | 1 | 0% | 2,732 | 3,822 | +40% | 0 | 0 | — |
case-16 | fail→fail | 14,387 | 23,402 | +63% | 1 | 1 | 0% | 2,057 | 3,942 | +92% | 0 | 0 | — |
case-17 | fail→fail | 13,852 | 17,527 | +27% | 1 | 1 | 0% | 1,550 | 3,770 | +143% | 0 | 0 | — |
case-18 | fail→pass | 14,920 | 18,736 | +26% | 1 | 1 | 0% | 2,566 | 3,957 | +54% | 0 | 0 | — |
case-19 | fail→fail | 27,928 | 29,316 | +5% | 1 | 1 | 0% | 3,595 | 5,023 | +40% | 0 | 0 | — |
case-20 | fail→fail | 16,005 | 19,725 | +23% | 1 | 1 | 0% | 2,619 | 3,980 | +52% | 0 | 0 | — |
case-21 | fail→pass | 13,834 | 20,334 | +47% | 1 | 1 | 0% | 2,273 | 4,225 | +86% | 0 | 0 | — |
case-22 | fail→fail | 13,002 | 28,165 | +117% | 1 | 1 | 0% | 2,360 | 4,516 | +91% | 0 | 0 | — |
case-23 | fail→fail | 11,676 | 27,257 | +133% | 1 | 1 | 0% | 2,082 | 4,211 | +102% | 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.
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.