Adonis0123/adonis-skills / skills/web-performance-audit
web-performance-audit
VerifiedRun a read-only, evidence-led runtime performance audit of a real web application, especially authenticated editors, dashboards, media tools, long-lived SPAs, and pages with many tabs or states. Use when the user asks to profile slowness, inspect every tab, use Chrome DevTools or Chrome Recorder, run React Scan, compare normal and throttled interactions, find memory or rendering risks, or produce a performance report and recommendations without changing code. Discover available browser capabilities, guide the user through official installation when an important capability is missing, route to validated fallbacks, isolate measurement pollution, and distinguish measured facts from screen observations, hypotheses, and unverified paths. Do not use for a Lighthouse-only SEO/accessibility audit, tool installation with no audit goal, or when the user already wants a specific optimization implemented.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 15.9 KB |
| agents/openai.yaml | config | 341 B |
| assets/demo-audit-ledger.json | config | 9.3 KB |
| assets/demo-audit-report.md | markdown | 10.4 KB |
| evals/benchmark-summary-2026-08-30.md | markdown | 2.0 KB |
| evals/evals.json | config | 8.1 KB |
| evals/files/ambiguous-metrics.json | config | 511 B |
| evals/files/dashboard-capabilities.json | config | 560 B |
| evals/files/video-editor-observations.json | config | 660 B |
| evals/host-acceptance-2026-08-30.md | markdown | 4.7 KB |
| evals/trigger-evals.json | config | 3.1 KB |
| references/capability-bootstrap.md | markdown | 10.0 KB |
| references/cross-agent-compatibility.md | markdown | 6.4 KB |
| references/ecosystem-comparison.md | markdown | 2.6 KB |
| references/measurement-playbook.md | markdown | 8.6 KB |
| references/report-contract.md | markdown | 8.2 KB |
| references/tool-routing.md | markdown | 9.9 KB |
| scripts/audit_ledger.py | script | 41.4 KB |
| scripts/test_audit_ledger.py | script | 16.9 KB |
| scripts/test_public_portability.py | script | 4.0 KB |
| showcase.yaml | config | 7.1 KB |
assets/demo-audit-report.md
View on GitHub# Synthetic editor performance audit demo
## Decision summary
- **F-001 · P2 · VERIFIED** — Synthetic playback switch records a 296 ms lab click duration: The fixture records a 296 ms lab interaction; it does not establish a product budget violation, regression, or visible delay.
- **F-002 · P2 · INFERENCE** — Offscreen mounted originals are a decode-pressure hypothesis: If the mounted originals are loaded and decoded, they may compete with scrolling and playback resources.
- **F-003 · P3 · VERIFIED** — The fixture contains repeated third-party requests: No user-visible performance impact is established by request count alone.
- Counterevidence (VERIFIED): The synthetic 10-cycle fixture does not show monotonic post-recovery DOM growth.
## Target and conditions
| Field | Value |
| ------------------ | --------------------------------------------------------------------------------------------- |
| URL | https://app.example.test/editor/sanitized-project |
| URL identity-safe | yes; Synthetic route; no account, query, fragment, or private project identifier is recorded. |
| Scope | Synthetic fixture demonstrating the report contract; not a live product claim |
| Browser/build | Sanitized Chrome trace fixture; development |
| Viewport | 1440 x 900; DPR 2 |
| Initial state | Paused on Media; instrumentation off |
| Write boundary | read-only |
| Requested surfaces | Media / Video, Library / All |
## Trust and privacy
| Field | Value |
| -------------------------------------- | ---------------------------------------------------------------------------------------- |
| Content trust | trusted |
| Sensitive data | absent |
| Browser profile | isolated |
| Browser tooling / authorization / risk | False / False / not_applicable |
| Target class | internal |
| CrUX / usage statistics / headers | not_applicable / not_applicable / redacted |
| Notes | Synthetic fixture contains no real URL, account, token, customer data, or response body. |
## Tool and noise ledger
| Tool/capability | Version | Status | Role | Proof | Caveat |
| ----------------------- | ------- | --------- | ------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- |
| Sanitized trace fixture | 1 | available | Demonstrate trace, DOM, network, and recovery evidence fields | discovery: assets/demo-audit-ledger.json loaded; handshake: schema_version 1.1 recognized; call: validator and renderer completed without product access | Synthetic values demonstrate output semantics and are not current product measurements. |
## Coverage
| Surface/flow | State | Conditions | Status | Evidence/notes |
| ------------- | --------------------------- | ------------------------------------------------------- | ------- | -------------------------- |
| Media / Video | cold first-open | synthetic fixture; normal CPU; instrumentation off | covered | TRACE-DEMO-000 |
| Media / Video | warm switch during playback | synthetic fixture; 4× CPU slowdown; instrumentation off | covered | TRACE-DEMO-001 |
| Library / All | cold first-open | synthetic 48-item fixture; normal CPU; paused | covered | DOM-DEMO-001; NET-DEMO-001 |
| Library / All | warm return | synthetic 48-item fixture; normal CPU; paused | covered | DOM-DEMO-002 |
| Library / All | 10-cycle recovery | synthetic 48-item fixture; normal CPU; fixed recovery | covered | MEM-DEMO-001 |
## Findings
### F-001 | P2 | VERIFIED | Synthetic playback switch records a 296 ms lab click duration
- **Surface:** Media / Video
- **Conditions:** synthetic fixture; playback active; 4× CPU slowdown; instrumentation off; three equivalent runs
- **User impact:** The fixture records a 296 ms lab interaction; it does not establish a product budget violation, regression, or visible delay.
- **Causal status:** measured
- **Caveats:** This is a lab click duration, not field INP and not a current product result.
| Signal | Value | Unit | Conditions | Source | Tool |
| --------------------- | ----: | ---- | ---------------------------------------------------------- | ------ | ----------------------- |
| click duration median | 296 | ms | three synthetic equivalent runs; playback; 4× CPU slowdown | trace | Sanitized trace fixture |
- **Recommendation:** Use a real trace to split renderer, media, framework, and third-party work before choosing an implementation.
- **Retest:** Repeat the same playback switch with normal CPU and 4× CPU slowdown, instrumentation off, after any authorized change.
- **Success criterion:** Meet a product-approved interaction budget with improvement larger than run-to-run variation and no media regression.
- **Stop condition:** Stop if a production-equivalent clean run does not reproduce the delay or the difference falls inside run-to-run variation.
### F-002 | P2 | INFERENCE | Offscreen mounted originals are a decode-pressure hypothesis
- **Surface:** Library / All
- **Conditions:** synthetic 48-item Library DOM inventory; 14 cards visible
- **User impact:** If the mounted originals are loaded and decoded, they may compete with scrolling and playback resources.
- **Causal status:** correlated
- **Caveats:** Mounted is not loaded, decoded, or resident; the nominal RGBA8 proxy is not JS heap or GPU residency.
| Signal | Value | Unit | Conditions | Source | Tool |
| ----------------------------------------- | ----: | -------- | --------------------------------------- | ------ | ----------------------- |
| offscreen mounted original-image elements | 19 | elements | synthetic 48-item Library DOM inventory | dom | Sanitized trace fixture |
- **Recommendation:** First verify loaded, decoded, and resident states plus interaction impact; only then consider card variants or windowing.
- **Retest:** Measure the same list with resource metadata, decoded-state evidence, a fixed scroll, and playback negative control.
- **Success criterion:** Demonstrate less decode/resource pressure and a material interaction improvement without blank cards.
- **Stop condition:** Stop if offscreen resources are not decoded/resident or the interaction signal does not improve.
### F-003 | P3 | VERIFIED | The fixture contains repeated third-party requests
- **Surface:** Library / All
- **Conditions:** synthetic tab-exploration window
- **User impact:** No user-visible performance impact is established by request count alone.
- **Causal status:** measured
- **Caveats:** Request count is not bandwidth, main-thread cost, privacy impact, or causal attribution.
| Signal | Value | Unit | Conditions | Source | Tool |
| ------------------------- | ----: | -------- | -------------------------------- | ------- | ----------------------- |
| third-party request count | 79 | requests | synthetic tab-exploration window | network | Sanitized trace fixture |
- **Recommendation:** Attribute transfer and main-thread work before changing tracker lifecycle.
- **Retest:** Compare an equivalent product-approved tracker-on/off or isolated initiator trace.
- **Success criterion:** Show a reproducible user-facing or resource improvement without losing required analytics.
- **Stop condition:** Stop if the requests are cached/idle and have no material transfer, task, or interaction cost.
## Negative findings
### VERIFIED | The synthetic 10-cycle fixture does not show monotonic post-recovery DOM growth.
- **Conditions:** ten equivalent cycles plus fixed recovery
- **Causal status:** measured
- **Scope limit:** Does not cover a long editing session or media decoder resources.
| Signal | Value | Unit | Conditions | Source | Tool |
| ---------------------------- | ----: | ----- | --------------------------- | -------------- | ----------------------- |
| post-recovery DOM node delta | -10 | nodes | synthetic ten-cycle fixture | browser_metric | Sanitized trace fixture |
## UNVERIFIED and production gaps
- **production-equivalent product behavior** — This artifact is a synthetic showcase, not a live browser run. Needed evidence: Run the same contract against the authorized exact target with clean production-equivalent traces.
## Restoration
- Final state: Paused on Media; instrumentation and throttle off
- Initial state restored: yes
- Instrumentation removed: yes
- Throttling removed: yes
- Media stopped: yes
- Product data changed: no
- Code changed: no
- Notes: Synthetic demo did not open or mutate a product page.
- Finished: 2026-08-30T00:05:00+00:00
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |