Skip to main content
skill CX Codex

network-request-auditing — Claude Code skill: what it does

Audits browser fetch and XHR calls after UI interactions to find errors, duplicate requests, and latency issues. Useful for API-heavy pages and backend changes.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add spencerpauly/awesome-cursor-skills --skill network-request-auditing -a claude-code
Open on GitHub

What network-request-auditing does

This skill inspects application network health by running browser_network_requests inside Cursor's integrated browser. The agent first navigates the page, clicks buttons, or submits forms to generate actual traffic. Once requests settle, it classifies every fetch and XHR call across failures, latency, duplicates, and security flaws.

The audit flags 4xx and 5xx statuses. It highlights duplicate calls triggered by React effects, oversized payloads, tokens exposed in query strings, and missing authentication headers. Preflight storms with excess OPTIONS requests also get flagged. After analysis, the agent outputs a structured issue breakdown with URL patterns and remediation steps.

The skill depends entirely on the cursor-ide-browser MCP tools to capture traffic. Network activity happening inside an iframe might not show up in the log, requiring manual inspection.

What it helps with

  • Detect duplicate fetch and XHR calls triggered by React effects
  • Flag slow API requests with server latency over 500 ms
  • Identify auth tokens or sensitive data leaked into query parameters
  • Catch CORS failures and 4xx or 5xx status codes after API updates

Example prompts for the agent

  • Navigate to the dashboard, trigger the table filters, and audit network requests for failures or duplicate calls
  • Audit API traffic on the checkout form and check for tokens leaked in query strings
  • Inspect fetch calls after submitting the form to find slow endpoints and missing auth headers

How to install network-request-auditing

1Skills CLI (Vercel)

npx skills add spencerpauly/awesome-cursor-skills --skill network-request-auditing -a claude-code

2Manually: put the files in place

~/.claude/skills/network-request-auditing/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/network-request-auditing/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about network-request-auditing

How do I install network-request-auditing in Claude Code?

Run npx skills add spencerpauly/awesome-cursor-skills --skill network-request-auditing -a claude-code in your terminal.

What does network-request-auditing do?

It audits fetch and XHR calls via browser_network_requests after UI actions. The skill reports 4xx and 5xx errors, duplicate endpoints, slow requests over 500 ms, CORS failures, and sensitive data exposed in query strings.

What dependencies are required to run this skill?

The skill requires the cursor-ide-browser MCP server to execute commands like browser_network_requests and browser_navigate. Note that network traffic inside iframe elements may not appear in the audit log.

Source text

SKILL.md · CC0-1.0
---
name: network-request-auditing
description: After navigating and interacting in Cursor's built-in browser, use browser_network_requests to audit every fetch/XHR for failures, slowness, duplicate calls, and suspicious payloads. Use for API-heavy pages and after backend or client networking changes.
user-invocable: true
---

# Network Request Auditing

Deep-dive **network** health using the `cursor-ide-browser` MCP. This skill focuses on `browser_network_requests` — not just “any 500s” but patterns that indicate bugs, waste, or security issues.

## How it works

1. Drive the app in the browser (navigate, click, submit forms) so real requests fire.
2. Call **`browser_network_requests`** after meaningful interactions (and after navigation settles).
3. Classify and report findings using the criteria below.

Follow `cursor-ide-browser` workflow rules: use `browser_snapshot` before structural interactions; after actions that change the page, take a fresh snapshot before the next interaction.

## Audit checklist

### Failures

- **4xx / 5xx** — list method, URL (path + query), status, and whether the UI handled the error.
- **CORS or network errors** — often misconfigured origins or mixed content.

### Performance

- **Slow requests** — flag requests with high latency (e.g. > 500 ms server time if timings are visible; otherwise note unusually large waterfalls).
- **Duplicate calls** — same URL + method fired multiple times in one user action (often a React effect or missing deduplication).
- **Oversized payloads** — responses that look huge for what the UI needs (suggest pagination, field selection, or compression).

### Security and privacy

- **Sensitive data in URLs** — tokens or PII in query strings.
- **Missing auth** — API calls that should send credentials or bearer tokens but do not (compare with adjacent authenticated calls).

### Correctness

- **Unexpected hosts** — calls to third parties not documented for the feature (trackers, accidental leaks).
- **Preflight storms** — excessive OPTIONS requests may indicate wrong CORS caching or too many distinct origins.

## Steps

1. **Start from a clean navigation** — `browser_navigate` to the target URL (or use an existing tab via `browser_tabs`).

2. **Exercise the feature** — interactions that trigger API usage (filters, infinite scroll, form save, modal open).

3. **Fetch network log** — `browser_network_requests` after each logical step if the page does multiple round-trips.

4. **Report** — structured output:
   - Summary counts (failed, slow, duplicate groups).
   - Table or bullet list of issues with **URL pattern** (not necessarily full secrets), **status**, **category** (failure / perf / security / correctness).
   - Recommended next code changes or investigations.

## Notes

- Iframe traffic may not appear in the same log — note if the feature runs inside an iframe.
- Compare against **expected** API design; a 404 might be correct for “optional resource not found” if handled in UI.
- Pair with `browser_console_messages` for errors that do not surface as failed HTTP (e.g. parse errors after 200).

© spencerpauly · CC0-1.0 · spencerpauly/awesome-cursor-skills

Similar skills

All skills for Browser and scraping →

From the same set60

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor