Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Benchmark — Performance Baseline & Regression Detection · affaan-m | DecimalAI

Benchmark

SkillScore45Maintained

benchmark

Use this skill to measure performance baselines, detect regressions before/after PRs, and compare stack alternatives.

Testingby affaan-m

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

QA & Test Engineering pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/benchmark/SKILL.md

decimalai skills pull benchmark --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/benchmark/SKILL.md https://app.decimal.ai/s/benchmark/SKILL.md

Raw file: /s/benchmark/SKILL.md ↗ · machine-readable summary at /s/benchmark.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 55d ago

Yes — +41% lift+2% tokens0% turns45% of whole cases pass with skill (10/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-22	✗→✓	▲ Improved	—	—
case-11	✗→✓	▲ Improved	—	—
case-08	✗→✓	▲ Improved	—	—
case-07	✗→✓	▲ Improved	—	—
case-12	✗→✓	▲ Improved	—	—

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

affaan-m / ECC

skills/benchmark/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

Skill contents

SKILL.md

Benchmark — Performance Baseline & Regression Detection

When to Use

Before and after a PR to measure performance impact
Setting up performance baselines for a project
When users report "it feels slow"
Before a launch — ensure you meet performance targets
Comparing your stack against alternatives

How It Works

Mode 1: Page Performance

Measures real browser metrics via browser MCP:

1. Navigate to each target URL
2. Measure Core Web Vitals:
   - LCP (Largest Contentful Paint) — target < 2.5s
   - CLS (Cumulative Layout Shift) — target < 0.1
   - INP (Interaction to Next Paint) — target < 200ms
   - FCP (First Contentful Paint) — target < 1.8s
   - TTFB (Time to First Byte) — target < 800ms
3. Measure resource sizes:
   - Total page weight (target < 1MB)
   - JS bundle size (target < 200KB gzipped)
   - CSS size
   - Image weight
   - Third-party script weight
4. Count network requests
5. Check for render-blocking resources

Mode 2: API Performance

Benchmarks API endpoints:

1. Hit each endpoint 100 times
2. Measure: p50, p95, p99 latency
3. Track: response size, status codes
4. Test under load: 10 concurrent requests
5. Compare against SLA targets

Mode 3: Build Performance

Measures development feedback loop:

1. Cold build time
2. Hot reload time (HMR)
3. Test suite duration
4. TypeScript check time
5. Lint time
6. Docker build time

Mode 4: Before/After Comparison

Run before and after a change to measure impact:

/benchmark baseline    # saves current metrics
# ... make changes ...
/benchmark compare     # compares against baseline

Output:

| Metric | Before | After | Delta | Verdict |
|--------|--------|-------|-------|---------|
| LCP | 1.2s | 1.4s | +200ms | WARNING: WARN |
| Bundle | 180KB | 175KB | -5KB | ✓ BETTER |
| Build | 12s | 14s | +2s | WARNING: WARN |

Output

Stores baselines in .ecc/benchmarks/ as JSON. Git-tracked so the team shares baselines.

Integration

CI: run /benchmark compare on every PR
Pair with /canary-watch for post-deploy monitoring
Pair with /browser-qa for full pre-ship checklist

Illustrative Want to see what benchmark would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/27/2026, 16:12 UTC · 22 cases

Lift

+41%

45% vs 5%

Agent turns

0%

1.0 vs 1.0

Tokens

+2%

2,118.82 vs 2,078.55

fail → pass · 9pass → pass · 1fail → fail · 12

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-22 We need to run a standard API baseline check against our `/api/v1/search` route. Many engineers default to 1,000 virtual users, but what is the standard concurrency parameter for this benchmarking process?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-06 We are measuring the Largest Contentful Paint for our web application landing page. Assuming standard performance baseline limits, what maximum threshold time should trigger a performance concern?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 Besides latency percentiles, what additional response characteristics must be captured when benchmarking backend REST endpoints?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-04 Before we ship our e-commerce checkout flow to production, we need a complete functional pre-ship QA audit verifying button clicks, modal interactions, form validation error messages, and cross-browser visual fidelity across mobile and desktop viewports.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-05 Our platform needs a 24-hour continuous distributed soak test simulating 50,000 requests per second across 5 global AWS regions to detect memory leaks under sustained endurance load.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 When generating and storing baseline performance artifacts so that the entire engineering team can track changes in version control, where should those benchmark JSON files be saved?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-08 When evaluating user interface responsiveness metrics for our web portal, what target limit should we enforce for Interaction to Next Paint?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-16 When auditing frontend assets beyond total bundle size on https://portal.example.com, what specific asset weight categories must be evaluated?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 We want to evaluate our application's compilation feedback loop after upgrading our bundler to Vite 5. Which benchmarking mode is appropriate for this task?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-07 Our design team added animated banners to https://app.example.com and we want to verify layout stability metrics. What is the target limit for Cumulative Layout Shift?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-12 We are about to refactor our database query layer. How should we record our baseline before changes and run the evaluation comparison after changes are completed?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-09 We need to set automated alerts for initial rendering speed on our marketing page. What maximum threshold should be enforced for First Contentful Paint?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-01 We are getting ready to launch our new landing page at https://example.com/dashboard and want to make sure it loads quickly. Could you analyze its core web vitals, bundle sizes, and network request counts to verify if the site meets recommended performance baselines?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-02 Our team suspects our backend services are slowing down under heavy traffic. Please run a performance evaluation on our `/api/v1/checkout` endpoint to collect latency percentiles and check its stability when handling concurrent requests.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-19 When rendering the output table comparing baseline metrics against post-change metrics, how should metric degradation or improvement be represented?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 Our developer experience team wants to benchmark developer feedback loop efficiency across our local workspace. Which build pipeline components should be measured?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-17 During a page performance evaluation of our web app, what structural assets check should be performed regarding network request order?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-10 Our frontend team wants to audit initial server responsiveness for dynamic HTTP requests. What baseline target maximum should be established for Time to First Byte?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 We deployed revision v2.4.1 to production 10 minutes ago. Please set up continuous automated post-deployment monitoring to watch live server error rates and alert on production anomalies over the next 4 hours.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 What automated command step should be integrated into our GitHub Actions continuous integration workflow on pull requests to prevent performance regressions?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-18 What HTTP connection metric should be tallied when evaluating page load overhead for a target URL?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-20 In what data format should baseline performance records be saved, and how should they be shared across the development team?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

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. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 4.6%
Pass rate with the skill: 45.5%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v1
Run completed: 7/27/2026, 16:12 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run45

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns-24% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 55 days ago

Other measured skills in the registry, with their headline benchmark lift.