Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Prompt for creating detailed feature implementation plans, following Epoch monorepo structure.
.claude/skills/breakdown-feature-implementation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 82% | 0% |
Act as an industry-veteran software engineer responsible for crafting high-touch features for large-scale SaaS companies. Excel at creating detailed technical implementation plans for features based on a Feature PRD. Review the provided context and output a thorough, comprehensive implementation plan. Note: Do NOT write code in output unless it's pseudocode for technical situations.
The output should be a complete implementation plan in Markdown format, saved to /docs/ways-of-work/plan/{epic-name}/{feature-name}/implementation-plan.md.
Folder and file structure for both front-end and back-end repositories following Epoch's monorepo structure:
apps/
[app-name]/
services/
[service-name]/
packages/
[package-name]/For each feature:
Feature goal described (3-5 sentences)
##### System Architecture Overview
Create a comprehensive system architecture diagram using Mermaid that shows how this feature integrates into the overall system. The diagram should include:
Use subgraphs to organize these layers clearly. Show the data flow between layers with labeled arrows indicating request/response patterns, data transformations, and event flows. Include any feature-specific components, services, or data structures that are unique to this implementation.
##### Database Schema Design
Create an entity-relationship diagram using Mermaid showing the feature's data model:
##### API Design
##### Frontend Architecture
###### Component Hierarchy Documentation
The component structure will leverage the shadcn/ui library for a consistent and accessible foundation.
Layout Structure:
Recipe Library Page
├── Header Section (shadcn: Card)
│ ├── Title (shadcn: Typography `h1`)
│ ├── Add Recipe Button (shadcn: Button with DropdownMenu)
│ │ ├── Manual Entry (DropdownMenuItem)
│ │ ├── Import from URL (DropdownMenuItem)
│ │ └── Import from PDF (DropdownMenuItem)
│ └── Search Input (shadcn: Input with icon)
├── Main Content Area (flex container)
│ ├── Filter Sidebar (aside)
│ │ ├── Filter Title (shadcn: Typography `h4`)
│ │ ├── Category Filters (shadcn: Checkbox group)
│ │ ├── Cuisine Filters (shadcn: Checkbox group)
│ │ └── Difficulty Filters (shadcn: RadioGroup)
│ └── Recipe Grid (main)
│ └── Recipe Card (shadcn: Card)
│ ├── Recipe Image (img)
│ ├── Recipe Title (shadcn: Typography `h3`)
│ ├── Recipe Tags (shadcn: Badge)
│ └── Quick Actions (shadcn: Button - View, Edit)##### Security Performance
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→fail | 21,946 | 26,659 | +21% | 1 | 1 | 0% | 3,623 | 5,820 | +61% | 0 | 0 | — |
case-01 | fail→fail | 31,251 | 34,303 | +10% | 1 | 1 | 0% | 6,242 | 7,181 | +15% | 0 | 0 | — |
case-02 | fail→pass | 26,478 | 30,619 | +16% | 1 | 1 | 0% | 4,964 | 6,761 | +36% | 0 | 0 | — |
case-03 | fail→pass | 30,478 | 31,301 | +3% | 1 | 1 | 0% | 6,240 | 7,179 | +15% | 0 | 0 | — |
case-04 | pass→fail | 21,685 | 32,238 | +49% | 1 | 1 | 0% | 3,509 | 7,128 | +103% | 0 | 0 | — |
case-05 | pass→fail | 23,325 | 34,228 | +47% | 1 | 1 | 0% | 4,976 | 7,110 | +43% | 0 | 0 | — |
case-06 | pass→fail | 17,281 | 35,147 | +103% | 1 | 1 | 0% | 2,766 | 7,098 | +157% | 0 | 0 | — |
case-08 | fail→pass | 22,074 | 31,865 | +44% | 1 | 1 | 0% | 3,534 | 6,617 | +87% | 0 | 0 | — |
case-09 | fail→pass | 18,977 | 23,634 | +25% | 1 | 1 | 0% | 3,250 | 5,090 | +57% | 0 | 0 | — |
case-10 | fail→pass | 19,617 | 25,501 | +30% | 1 | 1 | 0% | 3,106 | 5,639 | +82% | 0 | 0 | — |
case-11 | pass→pass | 19,954 | 33,483 | +68% | 1 | 1 | 0% | 3,121 | 7,116 | +128% | 0 | 0 | — |
case-12 | pass→pass | 18,767 | 24,714 | +32% | 1 | 1 | 0% | 3,613 | 5,203 | +44% | 0 | 0 | — |
case-13 | pass→pass | 24,432 | 34,952 | +43% | 1 | 1 | 0% | 3,611 | 7,109 | +97% | 0 | 0 | — |
case-14 | fail→fail | 32,768 | 31,462 | -4% | 1 | 1 | 0% | 6,178 | 7,117 | +15% | 0 | 0 | — |
case-15 | fail→fail | 22,685 | 23,795 | +5% | 1 | 1 | 0% | 3,946 | 5,210 | +32% | 0 | 0 | — |
case-16 | pass→pass | 13,277 | 22,432 | +69% | 1 | 1 | 0% | 2,045 | 4,906 | +140% | 0 | 0 | — |
case-17 | pass→pass | 21,133 | 34,965 | +65% | 1 | 1 | 0% | 3,493 | 7,116 | +104% | 0 | 0 | — |
case-18 | fail→pass | 15,451 | 25,712 | +66% | 1 | 1 | 0% | 2,454 | 5,626 | +129% | 0 | 0 | — |
case-19 | pass→pass | 16,132 | 29,688 | +84% | 1 | 1 | 0% | 2,421 | 3,754 | +55% | 0 | 0 | — |
case-20 | pass→pass | 28,491 | 35,448 | +24% | 1 | 1 | 0% | 4,680 | 7,111 | +52% | 0 | 0 | — |
case-21 | fail→fail | 26,677 | 32,147 | +21% | 1 | 1 | 0% | 4,151 | 6,791 | +64% | 0 | 0 | — |
case-22 | fail→fail | 19,621 | 21,390 | +9% | 1 | 1 | 0% | 3,194 | 5,016 | +57% | 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 +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/23/2026 | +50% |
Other measured skills in the registry, with their headline benchmark lift.