Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Estimate AI-assisted and hybrid human+agent development work with research-backed PERT statistics and calibration feedback loops
.claude/skills/sickn33-progressive-estimation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 38% | 0% |
Estimate AI-assisted and hybrid human+agent development work using research-backed formulas with PERT statistics, confidence bands, and calibration feedback loops.
Progressive Estimation adapts to your team's working mode — human-only, hybrid, or agent-first — applying the right velocity model and multipliers for each. It produces statistical estimates rather than gut feelings.
Single task: > "Estimate building a REST API with authentication using Claude Code"
Batch mode: > "Estimate these 12 JIRA tickets for our next sprint"
With context: > "We have 3 developers using AI agents for ~60% of implementation. Estimate this feature."
Solution: Use P75 or P90 for commitments, not P50
Solution: The skill asks clarifying questions — provide team size and agent usage
Solution: Re-calibrate when team composition or tooling changes significantly
@sprint-planning - Sprint planning and backlog management@project-management - General project management workflows@capacity-planning - Team velocity and capacity planning| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→pass | 7,917 | 7,430 | -6% | 1 | 1 | 0% | 1,504 | 2,077 | +38% | 0 | 0 | — |
case-02 | fail→fail | 11,728 | 16,732 | +43% | 1 | 1 | 0% | 2,479 | 4,425 | +78% | 0 | 0 | — |
case-03 | fail→pass | 16,616 | 11,951 | -28% | 1 | 1 | 0% | 3,171 | 2,839 | -10% | 0 | 0 | — |
case-01 | fail→fail | 12,574 | 18,259 | +45% | 1 | 1 | 0% | 2,359 | 3,913 | +66% | 0 | 0 | — |
case-04 | fail→fail | 15,239 | 18,400 | +21% | 1 | 1 | 0% | 2,933 | 4,605 | +57% | 0 | 0 | — |
case-05 | fail→pass | 16,054 | 16,653 | +4% | 1 | 1 | 0% | 3,123 | 4,238 | +36% | 0 | 0 | — |
case-06 | fail→pass | 10,629 | 12,381 | +16% | 1 | 1 | 0% | 1,935 | 3,333 | +72% | 0 | 0 | — |
case-07 | fail→pass | 14,364 | 13,862 | -3% | 1 | 1 | 0% | 2,432 | 3,364 | +38% | 0 | 0 | — |
case-08 | pass→pass | 10,191 | 8,363 | -18% | 1 | 1 | 0% | 1,873 | 2,387 | +27% | 0 | 0 | — |
case-09 | fail→fail | 10,138 | 7,875 | -22% | 1 | 1 | 0% | 1,850 | 2,179 | +18% | 0 | 0 | — |
case-11 | fail→pass | 12,463 | 11,748 | -6% | 1 | 1 | 0% | 2,094 | 2,848 | +36% | 0 | 0 | — |
case-12 | fail→pass | 14,290 | 10,773 | -25% | 1 | 1 | 0% | 2,476 | 2,719 | +10% | 0 | 0 | — |
case-13 | pass→pass | 19,038 | 16,209 | -15% | 1 | 1 | 0% | 3,640 | 4,270 | +17% | 0 | 0 | — |
case-14 | fail→pass | 13,663 | 12,094 | -11% | 1 | 1 | 0% | 2,451 | 3,060 | +25% | 0 | 0 | — |
case-15 | pass→pass | 13,858 | 11,269 | -19% | 1 | 1 | 0% | 2,491 | 2,700 | +8% | 0 | 0 | — |
case-16 | fail→pass | 12,098 | 13,693 | +13% | 1 | 1 | 0% | 2,139 | 3,005 | +40% | 0 | 0 | — |
case-17 | pass→pass | 13,189 | 10,421 | -21% | 1 | 1 | 0% | 2,162 | 2,566 | +19% | 0 | 0 | — |
case-18 | pass→pass | 15,239 | 14,625 | -4% | 1 | 1 | 0% | 2,729 | 3,504 | +28% | 0 | 0 | — |
case-19 | pass→pass | 3,732 | 4,594 | +23% | 1 | 1 | 0% | 858 | 1,756 | +105% | 0 | 0 | — |
case-20 | pass→pass | 15,402 | 9,365 | -39% | 1 | 1 | 0% | 2,737 | 2,441 | -11% | 0 | 0 | — |
case-21 | pass→pass | 14,585 | 12,165 | -17% | 1 | 1 | 0% | 2,381 | 2,806 | +18% | 0 | 0 | — |
case-22 | pass→pass | 14,031 | 9,870 | -30% | 1 | 1 | 0% | 3,087 | 3,049 | -1% | 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 +41 percentage points is the difference between those two pass rates over the 22 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.