Skip to content

Adonis0123/adonis-skills / skills/web-performance-audit

web-performance-audit

Verified

Run 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.

assets/demo-audit-ledger.json

View on GitHub
{
  "schema_version": "1.1",
  "audit": {
    "title": "Synthetic editor performance audit demo",
    "target_url": "https://app.example.test/editor/sanitized-project",
    "target_url_identity_safe": true,
    "target_url_redaction_notes": "Synthetic route; no account, query, fragment, or private project identifier is recorded.",
    "scope": "Synthetic fixture demonstrating the report contract; not a live product claim",
    "requested_surfaces": [
      {
        "id": "media-video",
        "label": "Media / Video",
        "required_states": ["cold first-open", "warm switch during playback"]
      },
      {
        "id": "library-all",
        "label": "Library / All",
        "required_states": [
          "cold first-open",
          "warm return",
          "10-cycle recovery"
        ]
      }
    ],
    "build_mode": "development",
    "browser": "Sanitized Chrome trace fixture",
    "viewport": "1440 x 900; DPR 2",
    "initial_state": "Paused on Media; instrumentation off",
    "restored_state": "Paused on Media; instrumentation and throttle off",
    "write_boundary": "read-only",
    "privacy": {
      "content_trust": "trusted",
      "sensitive_data": "absent",
      "browser_profile": "isolated",
      "performance_crux": "not_applicable",
      "usage_statistics": "not_applicable",
      "network_headers": "redacted",
      "browser_tooling_used": false,
      "browser_access_authorized": false,
      "target_class": "internal",
      "risk_decision": "not_applicable",
      "notes": "Synthetic fixture contains no real URL, account, token, customer data, or response body."
    },
    "started_at": "2026-08-30T00:00:00+00:00",
    "finished_at": "2026-08-30T00:05:00+00:00",
    "tools": [
      {
        "name": "Sanitized trace fixture",
        "version": "1",
        "status": "available",
        "role": "Demonstrate trace, DOM, network, and recovery evidence fields",
        "caveat": "Synthetic values demonstrate output semantics and are not current product measurements.",
        "proof": {
          "discovery": {
            "performed": true,
            "evidence": "assets/demo-audit-ledger.json loaded"
          },
          "handshake": {
            "performed": true,
            "evidence": "schema_version 1.1 recognized"
          },
          "harmless_call": {
            "performed": true,
            "evidence": "validator and renderer completed without product access"
          }
        }
      }
    ],
    "restoration": {
      "initial_state_restored": true,
      "instrumentation_removed": true,
      "throttling_removed": true,
      "media_stopped": true,
      "product_data_changed": false,
      "code_changed": false,
      "notes": "Synthetic demo did not open or mutate a product page."
    }
  },
  "coverage": [
    {
      "surface_id": "media-video",
      "surface": "Media / Video",
      "state": "cold first-open",
      "conditions": "synthetic fixture; normal CPU; instrumentation off",
      "status": "covered",
      "evidence": ["TRACE-DEMO-000"],
      "notes": ""
    },
    {
      "surface_id": "media-video",
      "surface": "Media / Video",
      "state": "warm switch during playback",
      "conditions": "synthetic fixture; 4× CPU slowdown; instrumentation off",
      "status": "covered",
      "evidence": ["TRACE-DEMO-001"],
      "notes": ""
    },
    {
      "surface_id": "library-all",
      "surface": "Library / All",
      "state": "cold first-open",
      "conditions": "synthetic 48-item fixture; normal CPU; paused",
      "status": "covered",
      "evidence": ["DOM-DEMO-001", "NET-DEMO-001"],
      "notes": ""
    },
    {
      "surface_id": "library-all",
      "surface": "Library / All",
      "state": "warm return",
      "conditions": "synthetic 48-item fixture; normal CPU; paused",
      "status": "covered",
      "evidence": ["DOM-DEMO-002"],
      "notes": ""
    },
    {
      "surface_id": "library-all",
      "surface": "Library / All",
      "state": "10-cycle recovery",
      "conditions": "synthetic 48-item fixture; normal CPU; fixed recovery",
      "status": "covered",
      "evidence": ["MEM-DEMO-001"],
      "notes": ""
    }
  ],
  "findings": [
    {
      "id": "F-001",
      "priority": "P2",
      "confidence": "VERIFIED",
      "title": "Synthetic playback switch records a 296 ms lab click duration",
      "surface_id": "media-video",
      "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.",
      "evidence": [
        {
          "signal": "click duration median",
          "value": "296",
          "unit": "ms",
          "conditions": "three synthetic equivalent runs; playback; 4× CPU slowdown",
          "source": "trace",
          "tool": "Sanitized trace fixture"
        }
      ],
      "recommendation": {
        "direction": "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."
      }
    },
    {
      "id": "F-002",
      "priority": "P2",
      "confidence": "INFERENCE",
      "title": "Offscreen mounted originals are a decode-pressure hypothesis",
      "surface_id": "library-all",
      "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.",
      "evidence": [
        {
          "signal": "offscreen mounted original-image elements",
          "value": "19",
          "unit": "elements",
          "conditions": "synthetic 48-item Library DOM inventory",
          "source": "dom",
          "tool": "Sanitized trace fixture"
        }
      ],
      "recommendation": {
        "direction": "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."
      }
    },
    {
      "id": "F-003",
      "priority": "P3",
      "confidence": "VERIFIED",
      "title": "The fixture contains repeated third-party requests",
      "surface_id": "library-all",
      "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.",
      "evidence": [
        {
          "signal": "third-party request count",
          "value": "79",
          "unit": "requests",
          "conditions": "synthetic tab-exploration window",
          "source": "network",
          "tool": "Sanitized trace fixture"
        }
      ],
      "recommendation": {
        "direction": "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": [
    {
      "confidence": "VERIFIED",
      "causal_status": "measured",
      "claim": "The synthetic 10-cycle fixture does not show monotonic post-recovery DOM growth.",
      "conditions": "ten equivalent cycles plus fixed recovery",
      "scope_limit": "Does not cover a long editing session or media decoder resources.",
      "evidence": [
        {
          "signal": "post-recovery DOM node delta",
          "value": "-10",
          "unit": "nodes",
          "conditions": "synthetic ten-cycle fixture",
          "source": "browser_metric",
          "tool": "Sanitized trace fixture"
        }
      ]
    }
  ],
  "unverified": [
    {
      "path": "production-equivalent product behavior",
      "reason": "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."
    }
  ]
}
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: 472kUpdated Claude CodeCodexVerifiedInstall
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: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall