Agent skill

add-report

Guide for adding new report cards to ClaudeBar that analyze local data sources and display metrics with comparison deltas. Use this skill when: (1) Adding a new report/analytics card (e.g., weekly summary, model breakdown, session stats) (2) Creating data analysis features that read local files and display aggregated metrics (3) Adding comparison cards that show "today vs previous" style deltas (4) Building any feature that follows the DailyUsage pattern (parse → aggregate → report → card)

Stars 938
Forks 77

Install this agent skill to your Project

npx add-skill https://github.com/tddworks/ClaudeBar/tree/main/.claude/skills/add-report

SKILL.md

Add Report Card to ClaudeBar

Add new report cards that analyze local data sources, compute metrics with comparison deltas, and display them in the existing card UI style using TDD.

When to Use

This skill covers adding report-style features — cards that:

  • Read and parse local data files (JSONL, JSON, CSV, etc.)
  • Aggregate metrics (cost, tokens, time, counts)
  • Compare periods (today vs yesterday, this week vs last week)
  • Display results in glassmorphism cards matching the existing UI

Architecture Pattern

Every report feature follows this data flow:

Data Source → Parser → Analyzer → Report Model → UsageSnapshot → Card View

Mapped to the codebase layers:

Layer Location What to Create
Domain Sources/Domain/{Feature}/ Rich models + @Mockable protocol
Infrastructure Sources/Infrastructure/{Provider}/ Parser + Analyzer implementation
App Sources/App/Views/ Card view(s)
Integration Provider class + statsGrid Wire analyzer → snapshot → UI

Reference implementation: See references/daily-usage-pattern.md for the complete DailyUsage feature as a working example of this pattern.

Workflow

Phase 0: Architecture Design (get user approval)
    ↓
Phase 1: Domain Models + Tests (TDD Red→Green)
    ↓
Phase 2: Infrastructure Parser + Analyzer + Tests
    ↓
Phase 3: Card View + Integration
    ↓
Phase 4: Verify all tests pass

Phase 0: Architecture Design (MANDATORY)

Before writing code, design the feature and get user approval.

Step 1: Define the Report

Answer these questions:

  • What data source? (JSONL sessions, API response, local DB, etc.)
  • What metrics? (cost, tokens, time, counts, etc.)
  • What comparison period? (today vs yesterday, this week vs last, etc.)
  • Which provider? (Claude, Codex, or generic across all?)
  • How many cards? (one per metric, or a single combined card?)

Step 2: Create Architecture Diagram

Example: Adding a weekly cost breakdown report

┌─────────────────────────────────────────────────────────┐
│  Data Source        Infrastructure          Domain       │
│                                                          │
│  ~/.claude/    →  WeeklyParser        →  WeeklyStat     │
│  projects/        (reads JSONL)          (per-day cost) │
│  *.jsonl                                                 │
│                    WeeklyAnalyzer     →  WeeklyReport   │
│                    (aggregates by        (this week vs  │
│                     week, implements      last week)    │
│                     protocol)                            │
│                          ↓                    ↓          │
│                    Provider.refresh()  → UsageSnapshot   │
│                                          .weeklyReport   │
│                          ↓                               │
│                    statsGrid() → WeeklyCardView          │
└─────────────────────────────────────────────────────────┘

Step 3: Document Components

Component Purpose Inputs Outputs
{Name}Stat Single period's data Raw aggregated values Formatted strings, isEmpty
{Name}Report Period comparison Two Stats Deltas, percentages, progress
{Name}Analyzing Protocol Date/config Report
{Name}Analyzer Implementation File paths Report
{Name}CardView UI card Report + metric Glassmorphism card

Step 4: Get User Approval

Use AskUserQuestion to confirm the design before proceeding.


Phase 1: Domain Models (TDD)

1a. Create the Stat Model

The stat model represents one period's aggregated data with rich formatting behavior.

Location: Sources/Domain/{Feature}/{Name}Stat.swift

Pattern to follow:

swift
import Foundation

public struct {Name}Stat: Sendable, Equatable {
    public let date: Date
    // Add your metrics here
    public let metricA: Decimal
    public let metricB: Int

    // Formatting — encapsulate in the model, not the view
    public var formattedMetricA: String { /* currency, compact number, duration, etc. */ }
    public var formattedMetricB: String { /* ... */ }

    public var isEmpty: Bool { /* all zeros check */ }
    public static func empty(for date: Date) -> Self { /* zero-valued instance */ }
}

Key rules:

  • Use Decimal for monetary values (not Double — floating point errors)
  • Use TimeInterval for durations
  • Use Locale(identifier: "en_US") for currency formatting (not en_US_POSIX which adds a space)
  • All formatting lives in the model — views just read formatted strings
  • isEmpty uses && (all zeros = empty), not ||

1b. Create the Report Model

The report model compares two periods and computes deltas.

Location: Sources/Domain/{Feature}/{Name}Report.swift

Pattern to follow:

swift
public struct {Name}Report: Sendable, Equatable {
    public let current: {Name}Stat   // e.g., today, this week
    public let previous: {Name}Stat  // e.g., yesterday, last week

    // Delta calculations
    public var metricADelta: Decimal { current.metricA - previous.metricA }
    public var metricAChangePercent: Double? {
        guard previous.metricA > 0 else { return nil }  // nil when previous is zero
        // ...
    }

    // Formatted deltas with sign: "+$5.00", "-1.2M"
    public var formattedMetricADelta: String { /* ... */ }

    // Progress for bar display (0-1 ratio of current vs total)
    public var metricAProgress: Double {
        let total = /* current + previous */
        guard total > 0 else { return 0 }
        return current / total
    }
}

Key rules:

  • Change percent returns nil when previous is zero (avoid division by zero)
  • Formatted deltas always include sign (+ or -)
  • Progress is current / (current + previous), clamped to 0-1
  • Use abs() for formatted values, prepend sign separately

1c. Create the Protocol

Location: Sources/Domain/{Feature}/{Name}Analyzing.swift

swift
import Mockable

@Mockable
public protocol {Name}Analyzing: Sendable {
    func analyze() async throws -> {Name}Report
}

1d. Write Tests First

Location: Tests/DomainTests/{Feature}/

Create two test files following Chicago School TDD (test state, not interactions):

  • {Name}StatTests.swift — Test formatting, isEmpty, edge cases
  • {Name}ReportTests.swift — Test deltas, percentages, nil cases, progress
swift
import Foundation
import Testing
@testable import Domain

@Suite
struct {Name}StatTests {
    @Test func `formats metric as expected`() {
        let stat = {Name}Stat(date: Date(), metricA: 14.26, ...)
        #expect(stat.formattedMetricA == "$14.26")
    }
    // ...
}

After writing tests → implement the models → run tests → all green.


Phase 2: Infrastructure (TDD)

2a. Create the Parser (if reading files)

If the report reads local files (JSONL, JSON, CSV), create a parser.

Location: Sources/Infrastructure/{Provider}/{Name}Parser.swift

Pattern: Parser is a struct (not protocol) since it's a pure data transformation.

swift
struct {Name}Parser {
    func parse(fileURL: URL) throws -> [{Name}Record] { /* ... */ }
    func parse(content: String) -> [{Name}Record] { /* for testing */ }
}

2b. Create the Analyzer

Location: Sources/Infrastructure/{Provider}/{Provider}{Name}Analyzer.swift

The analyzer implements the domain protocol and orchestrates:

  1. Find relevant files (filter by modification date for performance)
  2. Parse files into records
  3. Partition records by time period
  4. Aggregate into stat models
  5. Return report

Performance rule: Only scan files modified within the relevant time window. With 2000+ JSONL files, scanning all of them is too slow.

swift
public struct {Provider}{Name}Analyzer: {Name}Analyzing, Sendable {
    public func analyze() async throws -> {Name}Report {
        let files = findRecentFiles(since: periodStart)  // Performance!
        // parse → partition → aggregate → return
    }
}

2c. Write Infrastructure Tests

Location: Tests/InfrastructureTests/{Provider}/

  • Parser tests: valid input, missing fields, malformed data, timestamps
  • Analyzer tests: use temp directories with JSONL content, test date partitioning

Phase 3: Card View + Integration

3a. Create the Card View

Location: Sources/App/Views/{Name}CardView.swift

The card must match the existing glassmorphism style. Use WrappedStatCard as the reference:

swift
struct {Name}CardView: View {
    let metric: {Name}Metric  // enum for each displayable metric
    let report: {Name}Report
    let delay: Double          // for cascading entrance animation

    @Environment(\.appTheme) private var theme

    var body: some View {
        VStack(alignment: .leading, spacing: 6) {
            // 1. Header: icon + LABEL (uppercased)
            // 2. Large value (e.g., "$14.26" or "19.5M")
            // 3. Progress bar (animated)
            // 4. Delta comparison line (e.g., "Vs Mar 10 -$27.47 (4.9%)")
        }
        .padding(12)
        .background(/* theme.cardGradient + theme.glassBorder stroke */)
        .scaleEffect(isHovering ? 1.015 : 1.0)
        .onHover { isHovering = $0 }
    }
}

Card styling checklist:

  • .padding(12) on the VStack
  • theme.cardGradient fill + theme.glassBorder stroke (1pt)
  • theme.cardCornerRadius for corners
  • theme.fontDesign on all text
  • theme.textPrimary / theme.textSecondary / theme.textTertiary for text hierarchy
  • theme.progressTrack for bar background
  • Hover scale effect (1.015)
  • Animated progress bar with delay parameter

3b. Add to UsageSnapshot

Add an optional field for the report:

swift
// In Sources/Domain/Provider/UsageSnapshot.swift
public let {name}Report: {Name}Report?
// Add to init with default nil

3c. Wire into Provider

Inject the analyzer into the provider that owns this data:

swift
// In the provider's init:
private let {name}Analyzer: (any {Name}Analyzing)?

// In refresh():
snapshot = await attach{Name}Report(to: newSnapshot)

// Helper method:
private func attach{Name}Report(to snapshot: UsageSnapshot) async -> UsageSnapshot {
    guard let analyzer = {name}Analyzer,
          let report = try? await analyzer.analyze(),
          !report.current.isEmpty else { return snapshot }
    return UsageSnapshot(/* copy all fields, add report */)
}

3d. Render in statsGrid

Add to MenuContentView.statsGrid(snapshot:):

swift
if let report = snapshot.{name}Report {
    let baseDelay = Double(snapshot.quotas.count + 1) * 0.08
    // Render card(s) in LazyVGrid or standalone
}

3e. Register in ClaudeBarApp

Pass the analyzer when creating the provider:

swift
{Provider}Provider(
    probe: ...,
    settingsRepository: settingsRepository,
    {name}Analyzer: {Provider}{Name}Analyzer()
)

Phase 4: Verify

  1. tuist generate
  2. Run all tests: xcodebuild test -scheme ClaudeBar-Workspace ...
  3. Build the app and verify the cards appear
  4. Check logs for analyzer output

Checklist

Phase 0: Architecture

  • Define data source, metrics, comparison period, provider
  • Create architecture diagram
  • Get user approval

Phase 1: Domain (TDD)

  • Write {Name}StatTests (formatting, isEmpty, edge cases)
  • Implement {Name}Stat — make tests green
  • Write {Name}ReportTests (deltas, percentages, progress)
  • Implement {Name}Report — make tests green
  • Create {Name}Analyzing protocol with @Mockable

Phase 2: Infrastructure (TDD)

  • Write parser tests (valid input, missing fields, malformed data)
  • Implement parser — make tests green
  • Write analyzer tests (temp dirs, date partitioning)
  • Implement analyzer with file modification date filtering
  • Run all infrastructure tests green

Phase 3: Integration

  • Create {Name}CardView matching glassmorphism style
  • Add {name}Report field to UsageSnapshot
  • Wire analyzer into provider's refresh() via attach{Name}Report
  • Render cards in statsGrid
  • Register analyzer in ClaudeBarApp

Phase 4: Verify

  • tuist generate succeeds
  • All tests pass
  • App builds and cards render correctly
  • Logs show analyzer activity

Expand your agent's capabilities with these related and highly-rated skills.

tddworks/ClaudeBar

fix-bug

Guide for fixing bugs in ClaudeBar following Chicago School TDD and rich domain design. Use this skill when: (1) User reports a bug or unexpected behavior (2) Fixing a defect in existing functionality (3) User asks "fix this bug" or "this doesn't work correctly" (4) Correcting behavior that violates the user's mental model

938 77
Explore
tddworks/ClaudeBar

add-provider

Guide for adding new AI providers to ClaudeBar using TDD patterns. Use this skill when: (1) Adding a new AI assistant provider (like Antigravity, Cursor, etc.) (2) Creating a usage probe for a CLI tool or local API (3) Following TDD to implement provider integration (4) User asks "how do I add a new provider" or "create a provider for X"

938 77
Explore
tddworks/ClaudeBar

github-actions

Manage ClaudeBar's GitHub Actions CI/CD pipelines: build, test, and release workflows. Use this skill when: (1) Setting up secrets for CI/CD (certificate, API key, Sparkle key, Codecov) (2) Creating a new release — tag-based or manual workflow_dispatch (3) Triggering or explaining the build.yml, tests.yml, or release.yml workflows (4) Debugging release failures (signing, notarization, appcast) (5) Managing beta vs stable channels for Sparkle auto-updates (6) User says "release a new version", "push a tag", "set up CI secrets", "why did the release fail"

938 77
Explore
tddworks/ClaudeBar

implement-feature

Guide for implementing features in ClaudeBar following architecture-first design, TDD, rich domain models, and Swift 6.2 patterns. Use this skill when: (1) Adding new functionality to the app (2) Creating domain models that follow user's mental model (3) Building SwiftUI views that consume domain models directly (4) User asks "how do I implement X" or "add feature Y" (5) Implementing any feature that spans Domain, Infrastructure, and App layers

938 77
Explore
tddworks/ClaudeBar

improvement

Guide for making improvements to existing ClaudeBar functionality using TDD. Use this skill when: (1) Enhancing existing features (not adding new ones) (2) Improving UX, performance, or code quality (3) User asks "improve X", "make Y better", or "enhance Z" (4) Small enhancements that don't require full architecture design For NEW features, use implement-feature skill instead.

938 77
Explore
davila7/claude-code-templates

verl-rl-training

Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.

23,776 2,298
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results