Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Brief va quan ly agency, freelancer, vendor ben ngoai: scope of work, deliverable spec, revision rounds, quy trinh duyet, payment milestone, IP/bao mat, danh gia vendor sau project. Dung khi nhac 'brief agency', 'thue freelancer', 'quan ly vendor', 'scope of work', 'agency khong deliver dung', 'onboard vendor moi'.
.claude/skills/minhnv0807-67-agency-vendor-brief/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 15 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 98% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 159% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 58% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 67% | 0% |
> Agency lam sai thuong khong phai vi ho kem — ma vi brief te va ky vong khong khop tu dau. Brief tot voi agency: it revision, dung han, dung budget. Doi xu voi agency nhu doi tac chuyen nghiep, khong phai nguoi lam thue: brief ro + feedback ro = ket qua tot nhat.
Doc campaign brief (02-brief-chien-dich) va asset list (41-campaign-asset-list) neu co. Neu thieu, hoi toi da 4 cau:
Chot: muc tieu project (do duoc), budget tran, deadline cung, va deliverable toi thieu chap nhan duoc. Neu chua ro 4 thu nay, chua nen di brief agency.
Dung template o duoi. Gui kem brand guideline (46-brand-guideline), brand voice (35-brand-voice), va 3-5 reference tot + reference khong muon.
So sanh bao gia theo cung scope, khong so sanh theo tong tien. Chot lai bang van ban: scope, so vong revision, timeline, payment milestone.
Giu dau moi duy nhat hai ben. Feedback dung han da cam ket (agency tre vi minh feedback cham la loi minh). Feedback phai cu the va hanh dong duoc — khong noi "can hay hon".
Doi chieu deliverable spec. Duyet visual qua 47-design-review, duyet noi dung qua 62-marketing-review. Sau khi thanh toan xong, cham diem vendor va ghi vao database.
markdown# Agency/Vendor Brief — [Ten project] ## 1. Thong tin project | Hang muc | Chi tiet | |----------|----------| | Ten project | | | Ben thue | [Cong ty] — [Nguoi phu trach] — [Email/Phone] | | Vendor | [Ten] — [Contact person] — [Email] | | Loai cong viec | | | Ngay bat dau / deadline | | | Tong budget (da/chua VAT) | | ## 2. Boi canh va muc tieu Ve brand: [2-3 cau — brand la ai, dang lam gi, project nay fit vao chien luoc nao] Muc tieu project: [1 cau do duoc — vi du: "5 video UGC de test lam paid creative tren Meta, muc tieu CTR > 2%"] | Metric | Target | Cach do | Han | |--------|--------|---------|-----| ## 3. Scope of work | # | Deliverable | Spec chi tiet | Format output | So luong | |---|-------------|---------------|---------------|----------| Revision bao gom trong phi: [N] vong theo feedback. Vong bo sung: [phi]/vong. **KHONG bao gom:** [liet ke ro de chan scope creep] ## 4. Brand guideline phai tuan theo - Palette / typography / logo: [link brand kit] - Tone of voice: [link brand voice doc] - Reference tot: [3-5 link + ly do] - Reference KHONG lam theo: [link + ly do] ## 5. Deliverable spec | Deliverable | Platform | Size/Ratio | Thoi luong | Format file | Ngon ngu | |-------------|----------|------------|------------|-------------|----------| ## 6. Quy trinh duyet | Vong | Vendor nop | Han feedback tu minh | Ghi chu | |------|------------|----------------------|---------| | Draft | | +2 ngay lam viec | | | Revision | | +1 ngay lam viec | | | Final | | | Hard deadline | Tham quyen duyet: creative — leader; thay doi budget — CEO; gia han timeline — leader neu duoi 3 ngay, CEO neu tren 3 ngay. ## 7. Thanh toan | Milestone | % | So tien | Dieu kien | Thoi han chuyen | |-----------|---|---------|-----------|-----------------| | Ky hop dong | 30% | | Hop dong da ky | Truoc khi bat dau | | Duyet draft | 40% | | Draft duoc duyet | Trong 3 ngay | | Ban giao final | 30% | | Nghiem thu final | Trong 5 ngay | ## 8. Giao tiep Dau moi duy nhat: [ben minh] / [ben vendor]. Kenh chinh: [email/Zalo/Slack]. Nhip hop: [weekly / theo milestone]. SLA phan hoi cua minh: [N] ngay lam viec. ## 9. So huu tri tue va bao mat - Toan bo deliverable thuoc so huu ben thue sau khi thanh toan day du. - Vendor khong dang portfolio/social ve project khi chua co dong y bang van ban. - Thong tin project la bao mat, khong chia se ben thu ba. - Revision sau khi da nghiem thu final: tinh phi bo sung. ## 10. Danh gia vendor sau project | Tieu chi | Diem 1-5 | Ghi chu | |----------|----------|---------| | Chat luong output | | | | Dung han | | | | Giao tiep | | | | Toc do phan hoi | | | | Tuong xung gia tien | | | | **Tong the** | | | Tai su dung lan sau: Co / Khong — ly do.
02-brief-chien-dich: nguon context campaign cho phan boi canh va muc tieu.41-campaign-asset-list: danh sach deliverable can thue ngoai.61-budget-planning: xac dinh budget tran truoc khi di brief.47-design-review va 62-marketing-review: gate nghiem thu output.64-team-brief: dung khi nguoi nhan la nhan su noi bo thay vi vendor.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 34,694 | 31,631 | -9% | 1 | 1 | 0% | 5,192 | 6,650 | +28% | 0 | 0 | — |
case-02 | fail→fail | 35,793 | 26,744 | -25% | 1 | 1 | 0% | 5,377 | 6,660 | +24% | 0 | 0 | — |
case-03 | fail→fail | 33,810 | 30,858 | -9% | 1 | 1 | 0% | 4,864 | 6,976 | +43% | 0 | 0 | — |
case-04 | pass→pass | 24,886 | 16,981 | -32% | 1 | 1 | 0% | 3,047 | 4,663 | +53% | 0 | 0 | — |
case-05 | pass→pass | 18,455 | 22,288 | +21% | 1 | 1 | 0% | 3,043 | 5,359 | +76% | 0 | 0 | — |
case-06 | pass→pass | 26,396 | 21,664 | -18% | 1 | 1 | 0% | 3,317 | 5,806 | +75% | 0 | 0 | — |
case-07 | fail→pass | 20,621 | 25,565 | +24% | 1 | 1 | 0% | 2,666 | 5,269 | +98% | 0 | 0 | — |
case-08 | fail→pass | 23,544 | 19,060 | -19% | 1 | 1 | 0% | 2,590 | 4,020 | +55% | 0 | 0 | — |
case-09 | pass→pass | 19,900 | 20,143 | +1% | 1 | 1 | 0% | 2,495 | 4,360 | +75% | 0 | 0 | — |
case-10 | pass→pass | 21,755 | 21,521 | -1% | 1 | 1 | 0% | 2,262 | 4,532 | +100% | 0 | 0 | — |
case-11 | fail→pass | 19,923 | 26,124 | +31% | 1 | 1 | 0% | 2,372 | 6,137 | +159% | 0 | 0 | — |
case-12 | pass→pass | 22,842 | 22,012 | -4% | 1 | 1 | 0% | 2,993 | 4,790 | +60% | 0 | 0 | — |
case-13 | fail→fail | 15,291 | 13,357 | -13% | 1 | 1 | 0% | 2,479 | 4,203 | +70% | 0 | 0 | — |
case-14 | fail→pass | 15,196 | 9,012 | -41% | 1 | 1 | 0% | 1,658 | 2,616 | +58% | 0 | 0 | — |
case-15 | fail→pass | 22,747 | 24,918 | +10% | 1 | 1 | 0% | 2,899 | 4,850 | +67% | 0 | 0 | — |
case-16 | fail→pass | 19,244 | 20,650 | +7% | 1 | 1 | 0% | 3,051 | 5,217 | +71% | 0 | 0 | — |
case-17 | pass→pass | 19,440 | 15,075 | -22% | 1 | 1 | 0% | 2,408 | 4,207 | +75% | 0 | 0 | — |
case-18 | pass→pass | 20,792 | 16,932 | -19% | 1 | 1 | 0% | 2,132 | 4,474 | +110% | 0 | 0 | — |
case-19 | pass→pass | 23,149 | 16,908 | -27% | 1 | 1 | 0% | 2,849 | 4,667 | +64% | 0 | 0 | — |
case-20 | pass→pass | 19,669 | 12,131 | -38% | 1 | 1 | 0% | 2,418 | 3,898 | +61% | 0 | 0 | — |
case-21 | pass→pass | 19,786 | 20,605 | +4% | 1 | 1 | 0% | 2,099 | 4,107 | +96% | 0 | 0 | — |
case-22 | pass→pass | 19,132 | 20,110 | +5% | 1 | 1 | 0% | 2,149 | 4,435 | +106% | 0 | 0 | — |
case-23 | fail→fail | 32,163 | 23,029 | -28% | 1 | 1 | 0% | 2,717 | 4,804 | +77% | 0 | 0 | — |
case-24 | fail→pass | 12,776 | 9,387 | -27% | 1 | 1 | 0% | 1,133 | 2,726 | +141% | 0 | 0 | — |
case-25 | pass→pass | 21,302 | 24,333 | +14% | 1 | 1 | 0% | 2,493 | 4,738 | +90% | 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. 25 cases were attempted. The headline lift of +28 percentage points is the difference between those two pass rates over the 25 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.