# 📊 Agent Skill Benchmark Report > Generated: 2026-03-16T12:06:27.572Z > Token counting: `ceil(characters / 4)` — cl100k_base approximation. > Baselines: derived from **real, measured example prompts** (see Methodology). > Quality: structural rubric (0–10), no live LLM calls required. ## ❓ How to Read This Report This benchmark answers: **"How many tokens and dollars does an agent skill save compared to a developer writing the same guidance inline?"** **WITHOUT a skill**: A developer writes domain knowledge directly into the prompt every time (Baseline). **WITH a skill**: The agent loads the SKILL.md file (~400 tokens) — structured, reusable, cached. ## 🔢 Executive Summary | Metric | Value | | --------------------------------- | --------------------------------- | | Total Skills Benchmarked | **229** | | Avg. Tokens WITH Skill (SKILL.md) | **428 tokens** | | Baseline: Light prompt (no skill) | **1449 tokens** ↓ see Methodology | | Baseline: Heavy prompt (no skill) | **3656 tokens** ↓ see Methodology | | Avg. Token Savings vs Light | **70%** (1021 tokens/call) | | Avg. Token Savings vs Heavy | **88%** (3228 tokens/call) | | Avg. Quality Score | **9.9/10** | ## 📜 History | Version | Date | Skills | Avg Tokens | Savings (%) | Quality | Report | | ------- | ---------- | ------ | ---------- | ----------- | ------- | ------ | | v1.10.1 | 2026-03-16 | 229 | 428 | 88% | 9.9/10 | [Full Report](benchmarks/archive/v1.10.1.md) | | v1.10.0 | 2026-03-16 | 229 | 434 | 88% | 7/10 | [Full Report](benchmarks/archive/v1.10.0.md) | | v1.9.3 | 2026-03-15 | 229 | 460 | 87% | 8.9/10 | [Full Report](benchmarks/archive/v1.9.3.md) | | v1.9.2 | 2026-03-07 | 228 | 458 | 87% | 8.9/10 | [Full Report](benchmarks/archive/v1.9.2.md) | | v1.9.1 | 2026-03-07 | 228 | 458 | 87% | 8.9/10 | [Full Report](benchmarks/archive/v1.9.1.md) | | v1.9.0 | 2026-03-05 | 228 | 457 | 88% | 8.9/10 | [Full Report](benchmarks/archive/v1.9.0.md) | | v1.8.0 | 2026-03-02 | 228 | 443 | 88% | 8.9/10 | [Full Report](benchmarks/archive/v1.8.0.md) | | v1.7.3 | 2026-02-25 | 222 | 418 | 89% | 8.9/10 | [Full Report](benchmarks/archive/v1.7.3.md) | | v1.7.2 | 2026-02-25 | 220 | 413 | 89% | 8.9/10 | [Full Report](benchmarks/archive/v1.7.2.md) | ### 💰 Cost Comparison — Per Single Call (Average Skill) > Comparison based on **Heavy Baseline** vs. modern and speculative models. | Model | Original Cost | Skill Cost | Net Savings | % Saved | | ----------------- | ------------- | ---------- | -------------- | ------- | | Gemini 3 Flash | $0.0018280 | $0.0002140 | **$0.0016140 ** | 88% | | GPT-5 | $0.0045700 | $0.0005350 | **$0.0040350 ** | 88% | | Gemini 3.1 Pro | $0.0073120 | $0.0008560 | **$0.0064560 ** | 88% | | Claude Sonnet 4.5 | $0.0109680 | $0.0012840 | **$0.0096840 ** | 88% | ### 📈 Monthly Savings at Scale — (Avg Skill vs Heavy Prompt) | Daily Calls | Original Cost/mo | Monthly Savings (1 skill) | Monthly Savings (50 skills) | Model | | ----------- | ---------------- | ------------------------- | --------------------------- | ----- | | 1,000 | $137.1000 /mo | $121.0500 /mo | $6052.5000 /mo | GPT-5 | | 1,000 | $329.0400 /mo | $290.5200 /mo | $14526.0000 /mo | Claude Sonnet 4.5 | | 1,000 | $219.3600 /mo | $193.6800 /mo | $9684.0000 /mo | Gemini 3.1 Pro | ## 📦 Per-Category Summary

📦 android (22 skills | avg 290 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `android-architecture ` | 420 | █████████░ 89% | 10/10 | | `android-background-work` | 239 | █████████░ 93% | 10/10 | | `android-compose ` | 334 | █████████░ 91% | 10/10 | | `android-concurrency ` | 260 | █████████░ 93% | 10/10 | | `android-deployment ` | 287 | █████████░ 92% | 10/10 | | `android-design-system` | 426 | █████████░ 88% | 10/10 | | `android-di ` | 279 | █████████░ 92% | 10/10 | | `android-legacy-navigation` | 248 | █████████░ 93% | 10/10 | | `android-legacy-security` | 283 | █████████░ 92% | 10/10 | | `android-legacy-state ` | 251 | █████████░ 93% | 10/10 | | `android-navigation ` | 291 | █████████░ 92% | 10/10 | | `android-navigation-type-safe` | 257 | █████████░ 93% | 10/10 | | `android-networking ` | 303 | █████████░ 92% | 10/10 | | `android-notifications` | 310 | █████████░ 92% | 10/10 | | `android-performance ` | 291 | █████████░ 92% | 10/10 | | `android-persistence ` | 257 | █████████░ 93% | 10/10 | | `android-resources ` | 241 | █████████░ 93% | 10/10 | | `android-security ` | 332 | █████████░ 91% | 10/10 | | `android-state ` | 287 | █████████░ 92% | 10/10 | | `android-testing ` | 254 | █████████░ 93% | 10/10 | | `android-tooling ` | 246 | █████████░ 93% | 10/10 | | `android-xml-views ` | 287 | █████████░ 92% | 10/10 |

📦 angular (15 skills | avg 302 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `angular-architecture ` | 480 | █████████░ 87% | 10/10 | | `angular-component-patterns` | 311 | █████████░ 91% | 10/10 | | `angular-components ` | 344 | █████████░ 91% | 10/10 | | `angular-dependency-injection` | 293 | █████████░ 92% | 10/10 | | `angular-directives-pipes` | 257 | █████████░ 93% | 10/10 | | `angular-forms ` | 256 | █████████░ 93% | 10/10 | | `angular-http-client ` | 272 | █████████░ 93% | 10/10 | | `angular-performance ` | 295 | █████████░ 92% | 10/10 | | `angular-routing ` | 286 | █████████░ 92% | 10/10 | | `angular-rxjs-interop ` | 290 | █████████░ 92% | 10/10 | | `angular-security ` | 288 | █████████░ 92% | 10/10 | | `angular-ssr ` | 263 | █████████░ 93% | 10/10 | | `angular-state-management` | 292 | █████████░ 92% | 10/10 | | `angular-style-guide ` | 344 | █████████░ 91% | 10/10 | | `angular-testing ` | 260 | █████████░ 93% | 10/10 |

📦 common (25 skills | avg 535 tokens | quality 9.8/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `common-architecture-audit` | 700 | ████████░░ 81% | 10/10 | | `common-architecture-diagramming` | 399 | █████████░ 89% | 10/10 | | `common-best-practices` | 288 | █████████░ 92% | 10/10 | | `common-code-review ` | 385 | █████████░ 89% | 10/10 | | `common-context-optimization` | 471 | █████████░ 87% | 10/10 | | `common-debugging ` | 430 | █████████░ 88% | 10/10 | | `common-documentation ` | 564 | █████████░ 85% | 10/10 | | `common-error-handling` | 331 | █████████░ 91% | 10/10 | | `common-feedback-reporter` | 455 | █████████░ 88% | 10/10 | | `common-git-collaboration` | 675 | ████████░░ 82% | 10/10 | | `common-mobile-animation` | 365 | █████████░ 90% | 10/10 | | `common-mobile-ux-core` | 342 | █████████░ 91% | 10/10 | | `common-observability ` | 317 | █████████░ 91% | 10/10 | | `common-performance-engineering` | 686 | ████████░░ 81% | 10/10 | | `common-product-requirements` | 431 | █████████░ 88% | 10/10 | | `common-protocol-enforcement` | 415 | █████████░ 89% | 10/10 | | `common-security-audit` | 800 | ████████░░ 78% | 10/10 | | `common-security-standards` | 620 | ████████░░ 83% | 10/10 | | `common-session-retrospective` | 643 | ████████░░ 82% | 10/10 | | `common-skill-creator ` | 323 | █████████░ 91% | 10/10 | | `common-system-design ` | 532 | █████████░ 85% | 10/10 | | `common-tdd ` | 836 | ████████░░ 77% | 10/10 | | `common-workflow-writing` | 563 | █████████░ 85% | 10/10 | | `common-accessibility ` | 986 | ███████░░░ 73% | 8/10 | | `common-api-design ` | 819 | ████████░░ 78% | 8/10 |

📦 dart (3 skills | avg 460 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `dart-best-practices ` | 385 | █████████░ 89% | 10/10 | | `dart-language ` | 593 | ████████░░ 84% | 10/10 | | `dart-tooling ` | 403 | █████████░ 89% | 10/10 |

📦 database (3 skills | avg 524 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `database-mongodb ` | 588 | ████████░░ 84% | 10/10 | | `database-postgresql ` | 375 | █████████░ 90% | 10/10 | | `database-redis ` | 609 | ████████░░ 83% | 10/10 |

📦 flutter (21 skills | avg 453 tokens | quality 9.9/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `flutter-auto-route-navigation` | 427 | █████████░ 88% | 10/10 | | `flutter-bloc-state-management` | 417 | █████████░ 89% | 10/10 | | `flutter-cicd ` | 481 | █████████░ 87% | 10/10 | | `flutter-dependency-injection` | 406 | █████████░ 89% | 10/10 | | `flutter-design-system` | 525 | █████████░ 86% | 10/10 | | `flutter-error-handling` | 433 | █████████░ 88% | 10/10 | | `flutter-feature-based-clean-architecture` | 425 | █████████░ 88% | 10/10 | | `flutter-getx-navigation` | 382 | █████████░ 90% | 10/10 | | `flutter-go-router-navigation` | 440 | █████████░ 88% | 10/10 | | `flutter-idiomatic-flutter` | 371 | █████████░ 90% | 10/10 | | `flutter-layer-based-clean-architecture` | 510 | █████████░ 86% | 10/10 | | `flutter-localization ` | 494 | █████████░ 86% | 10/10 | | `flutter-navigation ` | 318 | █████████░ 91% | 10/10 | | `flutter-notifications` | 343 | █████████░ 91% | 10/10 | | `flutter-performance ` | 448 | █████████░ 88% | 10/10 | | `flutter-retrofit-networking` | 514 | █████████░ 86% | 10/10 | | `flutter-riverpod-state-management` | 444 | █████████░ 88% | 10/10 | | `flutter-security ` | 356 | █████████░ 90% | 10/10 | | `flutter-testing ` | 874 | ████████░░ 76% | 10/10 | | `flutter-widgets ` | 399 | █████████░ 89% | 10/10 | | `flutter-getx-state-management` | 503 | █████████░ 86% | 8/10 |

📦 golang (10 skills | avg 390 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `golang-api-server ` | 410 | █████████░ 89% | 10/10 | | `golang-architecture ` | 523 | █████████░ 86% | 10/10 | | `golang-concurrency ` | 401 | █████████░ 89% | 10/10 | | `golang-configuration ` | 307 | █████████░ 92% | 10/10 | | `golang-database ` | 360 | █████████░ 90% | 10/10 | | `golang-error-handling` | 320 | █████████░ 91% | 10/10 | | `golang-language ` | 324 | █████████░ 91% | 10/10 | | `golang-logging ` | 311 | █████████░ 91% | 10/10 | | `golang-security ` | 501 | █████████░ 86% | 10/10 | | `golang-testing ` | 443 | █████████░ 88% | 10/10 |

📦 ios (15 skills | avg 403 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `ios-app-lifecycle ` | 382 | █████████░ 90% | 10/10 | | `ios-architecture ` | 567 | ████████░░ 84% | 10/10 | | `ios-dependency-injection` | 392 | █████████░ 89% | 10/10 | | `ios-deployment ` | 362 | █████████░ 90% | 10/10 | | `ios-design-system ` | 391 | █████████░ 89% | 10/10 | | `ios-localization ` | 469 | █████████░ 87% | 10/10 | | `ios-navigation ` | 289 | █████████░ 92% | 10/10 | | `ios-networking ` | 442 | █████████░ 88% | 10/10 | | `ios-notifications ` | 304 | █████████░ 92% | 10/10 | | `ios-performance ` | 409 | █████████░ 89% | 10/10 | | `ios-persistence ` | 452 | █████████░ 88% | 10/10 | | `ios-security ` | 412 | █████████░ 89% | 10/10 | | `ios-state-management ` | 427 | █████████░ 88% | 10/10 | | `ios-swiftui ` | 310 | █████████░ 92% | 10/10 | | `ios-ui-navigation ` | 434 | █████████░ 88% | 10/10 |

📦 java (5 skills | avg 519 tokens | quality 9.6/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `java-concurrency ` | 470 | █████████░ 87% | 10/10 | | `java-language ` | 633 | ████████░░ 83% | 10/10 | | `java-testing ` | 523 | █████████░ 86% | 10/10 | | `java-tooling ` | 412 | █████████░ 89% | 10/10 | | `java-best-practices ` | 556 | █████████░ 85% | 8/10 |

📦 javascript (3 skills | avg 382 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `javascript-best-practices` | 403 | █████████░ 89% | 10/10 | | `javascript-language ` | 405 | █████████░ 89% | 10/10 | | `javascript-tooling ` | 339 | █████████░ 91% | 10/10 |

📦 kotlin (4 skills | avg 450 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `kotlin-best-practices` | 528 | █████████░ 86% | 10/10 | | `kotlin-coroutines ` | 330 | █████████░ 91% | 10/10 | | `kotlin-language ` | 551 | █████████░ 85% | 10/10 | | `kotlin-tooling ` | 392 | █████████░ 89% | 10/10 |

📦 laravel (10 skills | avg 360 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `laravel-api ` | 336 | █████████░ 91% | 10/10 | | `laravel-architecture ` | 375 | █████████░ 90% | 10/10 | | `laravel-background-processing` | 381 | █████████░ 90% | 10/10 | | `laravel-clean-architecture` | 428 | █████████░ 88% | 10/10 | | `laravel-database-expert` | 400 | █████████░ 89% | 10/10 | | `laravel-eloquent ` | 323 | █████████░ 91% | 10/10 | | `laravel-security ` | 334 | █████████░ 91% | 10/10 | | `laravel-sessions-middleware` | 382 | █████████░ 90% | 10/10 | | `laravel-testing ` | 313 | █████████░ 91% | 10/10 | | `laravel-tooling ` | 324 | █████████░ 91% | 10/10 |

📦 nestjs (21 skills | avg 579 tokens | quality 9.7/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `nestjs-api-standards ` | 516 | █████████░ 86% | 10/10 | | `nestjs-architecture ` | 406 | █████████░ 89% | 10/10 | | `nestjs-bullmq ` | 866 | ████████░░ 76% | 10/10 | | `nestjs-caching ` | 660 | ████████░░ 82% | 10/10 | | `nestjs-configuration ` | 562 | █████████░ 85% | 10/10 | | `nestjs-database ` | 628 | ████████░░ 83% | 10/10 | | `nestjs-deployment ` | 588 | ████████░░ 84% | 10/10 | | `nestjs-error-handling` | 532 | █████████░ 85% | 10/10 | | `nestjs-file-uploads ` | 366 | █████████░ 90% | 10/10 | | `nestjs-notification ` | 439 | █████████░ 88% | 10/10 | | `nestjs-observability ` | 607 | ████████░░ 83% | 10/10 | | `nestjs-real-time ` | 649 | ████████░░ 82% | 10/10 | | `nestjs-scheduling ` | 372 | █████████░ 90% | 10/10 | | `nestjs-search ` | 467 | █████████░ 87% | 10/10 | | `nestjs-security ` | 533 | █████████░ 85% | 10/10 | | `nestjs-security-isolation` | 538 | █████████░ 85% | 10/10 | | `nestjs-testing ` | 552 | █████████░ 85% | 10/10 | | `nestjs-transport ` | 475 | █████████░ 87% | 10/10 | | `nestjs-controllers-services` | 691 | ████████░░ 81% | 8/10 | | `nestjs-documentation ` | 872 | ████████░░ 76% | 8/10 | | `nestjs-performance ` | 831 | ████████░░ 77% | 8/10 |

📦 nextjs (18 skills | avg 465 tokens | quality 9.7/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `nextjs-app-router ` | 545 | █████████░ 85% | 10/10 | | `nextjs-architecture ` | 839 | ████████░░ 77% | 10/10 | | `nextjs-caching ` | 422 | █████████░ 88% | 10/10 | | `nextjs-data-access-layer` | 393 | █████████░ 89% | 10/10 | | `nextjs-data-fetching ` | 415 | █████████░ 89% | 10/10 | | `nextjs-i18n ` | 528 | █████████░ 86% | 10/10 | | `nextjs-pages-router ` | 338 | █████████░ 91% | 10/10 | | `nextjs-rendering ` | 398 | █████████░ 89% | 10/10 | | `nextjs-security ` | 333 | █████████░ 91% | 10/10 | | `nextjs-server-components` | 478 | █████████░ 87% | 10/10 | | `nextjs-state-management` | 574 | ████████░░ 84% | 10/10 | | `nextjs-styling ` | 456 | █████████░ 88% | 10/10 | | `nextjs-testing ` | 344 | █████████░ 91% | 10/10 | | `nextjs-tooling ` | 342 | █████████░ 91% | 10/10 | | `nextjs-upgrade ` | 497 | █████████░ 86% | 10/10 | | `nextjs-authentication` | 293 | █████████░ 92% | 8/10 | | `nextjs-optimization ` | 583 | ████████░░ 84% | 8/10 | | `nextjs-server-actions` | 586 | ████████░░ 84% | 8/10 |

📦 php (7 skills | avg 319 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `php-best-practices ` | 330 | █████████░ 91% | 10/10 | | `php-concurrency ` | 287 | █████████░ 92% | 10/10 | | `php-error-handling ` | 341 | █████████░ 91% | 10/10 | | `php-language ` | 326 | █████████░ 91% | 10/10 | | `php-security ` | 334 | █████████░ 91% | 10/10 | | `php-testing ` | 300 | █████████░ 92% | 10/10 | | `php-tooling ` | 316 | █████████░ 91% | 10/10 |

📦 quality-engineering (4 skills | avg 481 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `quality-engineering-business-analysis` | 423 | █████████░ 88% | 10/10 | | `quality-engineering-jira-integration` | 529 | █████████░ 86% | 10/10 | | `quality-engineering-quality-assurance` | 372 | █████████░ 90% | 10/10 | | `quality-engineering-zephyr-test-generation` | 601 | ████████░░ 84% | 10/10 |

📦 react (8 skills | avg 411 tokens | quality 9.8/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `react-component-patterns` | 407 | █████████░ 89% | 10/10 | | `react-hooks ` | 549 | █████████░ 85% | 10/10 | | `react-security ` | 342 | █████████░ 91% | 10/10 | | `react-state-management` | 390 | █████████░ 89% | 10/10 | | `react-testing ` | 326 | █████████░ 91% | 10/10 | | `react-tooling ` | 293 | █████████░ 92% | 10/10 | | `react-typescript ` | 333 | █████████░ 91% | 10/10 | | `react-performance ` | 648 | ████████░░ 82% | 8/10 |

📦 react-native (13 skills | avg 440 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `react-native-architecture` | 715 | ████████░░ 80% | 10/10 | | `react-native-components` | 501 | █████████░ 86% | 10/10 | | `react-native-deployment` | 444 | █████████░ 88% | 10/10 | | `react-native-dls ` | 258 | █████████░ 93% | 10/10 | | `react-native-navigation` | 299 | █████████░ 92% | 10/10 | | `react-native-navigation-v6` | 453 | █████████░ 88% | 10/10 | | `react-native-notifications` | 322 | █████████░ 91% | 10/10 | | `react-native-performance` | 501 | █████████░ 86% | 10/10 | | `react-native-platform-specific` | 425 | █████████░ 88% | 10/10 | | `react-native-security` | 523 | █████████░ 86% | 10/10 | | `react-native-state-management` | 419 | █████████░ 89% | 10/10 | | `react-native-styling ` | 411 | █████████░ 89% | 10/10 | | `react-native-testing ` | 445 | █████████░ 88% | 10/10 |

📦 spring-boot (10 skills | avg 370 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `spring-boot-api-design` | 309 | █████████░ 92% | 10/10 | | `spring-boot-architecture` | 622 | ████████░░ 83% | 10/10 | | `spring-boot-best-practices` | 387 | █████████░ 89% | 10/10 | | `spring-boot-data-access` | 316 | █████████░ 91% | 10/10 | | `spring-boot-deployment` | 333 | █████████░ 91% | 10/10 | | `spring-boot-microservices` | 315 | █████████░ 91% | 10/10 | | `spring-boot-observability` | 311 | █████████░ 91% | 10/10 | | `spring-boot-scheduling` | 268 | █████████░ 93% | 10/10 | | `spring-boot-security ` | 520 | █████████░ 86% | 10/10 | | `spring-boot-testing ` | 320 | █████████░ 91% | 10/10 |

📦 swift (8 skills | avg 353 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `swift-best-practices ` | 349 | █████████░ 90% | 10/10 | | `swift-concurrency ` | 341 | █████████░ 91% | 10/10 | | `swift-error-handling ` | 315 | █████████░ 91% | 10/10 | | `swift-language ` | 338 | █████████░ 91% | 10/10 | | `swift-memory-management` | 336 | █████████░ 91% | 10/10 | | `swift-swiftui ` | 380 | █████████░ 90% | 10/10 | | `swift-testing ` | 391 | █████████░ 89% | 10/10 | | `swift-tooling ` | 373 | █████████░ 90% | 10/10 |

📦 typescript (4 skills | avg 510 tokens | quality 10.0/10)

| Skill | Tokens | Savings (vs Heavy) | Quality | | ----------------------- | ------ | ------------------ | ------- | | `typescript-best-practices` | 483 | █████████░ 87% | 10/10 | | `typescript-language ` | 547 | █████████░ 85% | 10/10 | | `typescript-security ` | 425 | █████████░ 88% | 10/10 | | `typescript-tooling ` | 585 | ████████░░ 84% | 10/10 |
## 🏆 Quality Leaders | Rank | Skill | Category | Quality | Tokens | | ---- | ----------------------- | -------- | ------- | ------ | | 1 | `android-architecture ` | android | 10/10 | 420 | | 2 | `android-background-work` | android | 10/10 | 239 | | 3 | `android-compose ` | android | 10/10 | 334 | | 4 | `android-concurrency ` | android | 10/10 | 260 | | 5 | `android-deployment ` | android | 10/10 | 287 | | 6 | `android-design-system` | android | 10/10 | 426 | | 7 | `android-di ` | android | 10/10 | 279 | | 8 | `android-legacy-navigation` | android | 10/10 | 248 | | 9 | `android-legacy-security` | android | 10/10 | 283 | | 10 | `android-legacy-state ` | android | 10/10 | 251 | ## 📐 Methodology & Baseline Justification ### Why These Baselines? The baselines are derived from **real, token-counted example prompts** that represent what a developer actually writes when there is no structured skill available. Using NestJS as the **Reference Unit**: Because we measure instruction volume replaced, using a high-density reference ensures scientific consistency across all tech stacks. #### 🟡 Reference Technical Prompt — Light — 1449 tokens > **Reference Technical Prompt — Light (e.g., NestJS)** > A compact inline system prompt used as a reference for token count calibration. Representative of focused developer instructions without a structured skill. #### 🔴 Reference Technical Prompt — Heavy — 3656 tokens > **Reference Technical Prompt — Heavy (e.g., NestJS Architecture)** > A comprehensive architect-level inline prompt used as a reference for complex tasks. Includes deep patterns and rules sent by developers when no skill is present. ### 🏆 Detailed Quality Rubric (0–10) To ensure skills are not just "short" but actually **high quality**, every skill is scored against this structural rubric: | Score | Criteria | Rationale | | ------ | ------------------------- | ------------------------------------------------------ | | **+2** | **Structured Guidelines** | At least 3 specific instructions/bullet points. | | **+2** | **Anti-Patterns** | Specifically listing what the LLM should *avoid*. | | **+2** | **Reference Examples** | Presence of a verified `references/` folder with code. | | **+2** | **Token Optimality** | Entire `SKILL.md` is ≤100 lines (forces brevity). | | **+2** | **Trigger Metadata** | Proper keywords and file-match triggers defined. | ### 🛡️ How to Verify This Report Trust but verify. You can audit the raw data and run the benchmark yourself: 1. **Clone the repo** and install dependencies (`pnpm install`). 2. **Inspect Source**: The benchmark logic is open in [cli/src/scripts/benchmark/](./cli/src/scripts/benchmark/). ### Pricing (per 1M input tokens, Feb 2026) - **Gemini 3 Flash**: $0.50 - **GPT-5**: $1.25 - **Gemini 3.1 Pro**: $2.00 - **Claude Sonnet 4.5**: $3.00