Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Track construction submittals through the review process. Manage approvals, revisions, and compliance.
.claude/skills/datadrivenconstruction-submittal-tracker/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 74% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 20% | 0% |
Submittals require careful tracking to avoid delays. This skill manages the entire submittal workflow from creation to approval.
pythonimport pandas as pd from datetime import datetime, date, timedelta from typing import Dict, Any, List, Optional from dataclasses import dataclass, field from enum import Enum class SubmittalStatus(Enum): DRAFT = "draft" SUBMITTED = "submitted" UNDER_REVIEW = "under_review" APPROVED = "approved" APPROVED_AS_NOTED = "approved_as_noted" REVISE_RESUBMIT = "revise_resubmit" REJECTED = "rejected" class SubmittalType(Enum): PRODUCT_DATA = "product_data" SHOP_DRAWING = "shop_drawing" SAMPLE = "sample" MOCK_UP = "mock_up" CERTIFICATE = "certificate" TEST_REPORT = "test_report" @dataclass class ReviewComment: reviewer: str comment: str comment_date: date resolved: bool = False @dataclass class Submittal: submittal_id: str spec_section: str title: str submittal_type: SubmittalType status: SubmittalStatus contractor: str revision: int submitted_date: Optional[date] required_date: date approved_date: Optional[date] = None comments: List[ReviewComment] = field(default_factory=list) files: List[str] = field(default_factory=list) class SubmittalTracker: def __init__(self, project_name: str): self.project_name = project_name self.submittals: Dict[str, Submittal] = {} self._counter = 0 def create_submittal(self, spec_section: str, title: str, submittal_type: SubmittalType, contractor: str, required_date: date) -> Submittal: self._counter += 1 sub_id = f"SUB-{self._counter:04d}" submittal = Submittal( submittal_id=sub_id, spec_section=spec_section, title=title, submittal_type=submittal_type, status=SubmittalStatus.DRAFT, contractor=contractor, revision=0, submitted_date=None, required_date=required_date ) self.submittals[sub_id] = submittal return submittal def submit(self, sub_id: str, files: List[str]): if sub_id in self.submittals: self.submittals[sub_id].status = SubmittalStatus.SUBMITTED self.submittals[sub_id].submitted_date = date.today() self.submittals[sub_id].files = files def review(self, sub_id: str, status: SubmittalStatus, reviewer: str, comment: str = ""): if sub_id in self.submittals: sub = self.submittals[sub_id] sub.status = status if comment: sub.comments.append(ReviewComment(reviewer, comment, date.today())) if status in [SubmittalStatus.APPROVED, SubmittalStatus.APPROVED_AS_NOTED]: sub.approved_date = date.today() def resubmit(self, sub_id: str, files: List[str]): if sub_id in self.submittals: sub = self.submittals[sub_id] sub.revision += 1 sub.status = SubmittalStatus.SUBMITTED sub.submitted_date = date.today() sub.files = files def get_pending(self) -> List[Submittal]: return [s for s in self.submittals.values() if s.status in [SubmittalStatus.SUBMITTED, SubmittalStatus.UNDER_REVIEW]] def get_overdue(self) -> List[Submittal]: today = date.today() return [s for s in self.submittals.values() if s.status not in [SubmittalStatus.APPROVED, SubmittalStatus.APPROVED_AS_NOTED] and s.required_date < today] def export_log(self, output_path: str): data = [{ 'ID': s.submittal_id, 'Spec': s.spec_section, 'Title': s.title, 'Type': s.submittal_type.value, 'Status': s.status.value, 'Rev': s.revision, 'Submitted': s.submitted_date, 'Required': s.required_date, 'Approved': s.approved_date } for s in self.submittals.values()] pd.DataFrame(data).to_excel(output_path, index=False)
pythontracker = SubmittalTracker("Office Tower") sub = tracker.create_submittal("09 29 00", "Gypsum Board", SubmittalType.PRODUCT_DATA, "ABC Drywall", date.today() + timedelta(days=14)) tracker.submit(sub.submittal_id, ["spec_sheet.pdf"]) tracker.review(sub.submittal_id, SubmittalStatus.APPROVED, "Architect")
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→pass | 12,672 | 5,295 | -58% | 1 | 1 | 0% | 2,171 | 2,339 | +8% | 0 | 0 | — |
case-01 | fail→pass | 12,321 | 13,505 | +10% | 1 | 1 | 0% | 2,498 | 4,350 | +74% | 0 | 0 | — |
case-02 | fail→fail | 16,543 | 11,801 | -29% | 1 | 1 | 0% | 3,338 | 3,525 | +6% | 0 | 0 | — |
case-03 | fail→pass | 18,141 | 11,621 | -36% | 1 | 1 | 0% | 3,880 | 3,812 | -2% | 0 | 0 | — |
case-04 | pass→pass | 4,833 | 2,515 | -48% | 1 | 1 | 0% | 811 | 1,853 | +128% | 0 | 0 | — |
case-05 | fail→pass | 9,669 | 2,330 | -76% | 1 | 1 | 0% | 1,573 | 1,750 | +11% | 0 | 0 | — |
case-06 | pass→pass | 7,797 | 3,204 | -59% | 1 | 1 | 0% | 1,292 | 1,971 | +53% | 0 | 0 | — |
case-07 | fail→pass | 10,722 | 5,846 | -45% | 1 | 1 | 0% | 2,065 | 2,468 | +20% | 0 | 0 | — |
case-08 | pass→pass | 13,737 | 3,649 | -73% | 1 | 1 | 0% | 2,128 | 1,986 | -7% | 0 | 0 | — |
case-09 | fail→pass | 8,637 | 2,480 | -71% | 1 | 1 | 0% | 1,386 | 1,812 | +31% | 0 | 0 | — |
case-11 | pass→pass | 6,131 | 1,808 | -71% | 1 | 1 | 0% | 1,043 | 1,636 | +57% | 0 | 0 | — |
case-12 | fail→pass | 10,837 | 6,110 | -44% | 1 | 1 | 0% | 1,711 | 2,357 | +38% | 0 | 0 | — |
case-13 | pass→pass | 4,278 | 1,100 | -74% | 1 | 1 | 0% | 642 | 1,492 | +132% | 0 | 0 | — |
case-14 | fail→fail | 12,079 | 7,778 | -36% | 1 | 1 | 0% | 2,366 | 2,919 | +23% | 0 | 0 | — |
case-15 | pass→fail | 13,307 | 4,388 | -67% | 1 | 1 | 0% | 2,611 | 2,205 | -16% | 0 | 0 | — |
case-16 | pass→pass | 10,358 | 3,139 | -70% | 1 | 1 | 0% | 1,897 | 1,934 | +2% | 0 | 0 | — |
case-17 | pass→pass | 8,289 | 3,877 | -53% | 1 | 1 | 0% | 1,278 | 2,079 | +63% | 0 | 0 | — |
case-18 | fail→pass | 12,182 | 3,451 | -72% | 1 | 1 | 0% | 2,036 | 1,919 | -6% | 0 | 0 | — |
case-19 | fail→pass | 8,636 | 2,716 | -69% | 1 | 1 | 0% | 1,552 | 1,862 | +20% | 0 | 0 | — |
case-20 | pass→pass | 17,320 | 14,427 | -17% | 1 | 1 | 0% | 3,723 | 4,217 | +13% | 0 | 0 | — |
case-21 | pass→pass | 19,426 | 18,839 | -3% | 1 | 1 | 0% | 3,909 | 5,722 | +46% | 0 | 0 | — |
case-22 | pass→pass | 19,459 | 14,200 | -27% | 1 | 1 | 0% | 4,133 | 4,418 | +7% | 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. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 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.