generate() — Create Artifacts

The generate tool creates production-ready artifacts from captured browser data. Tests, reproduction scripts, reports, and exports — all generated from real browser sessions.

Quick Reference

generate({format: "test"})                               // Playwright regression test
generate({format: "reproduction"})                        // Bug reproduction script
generate({format: "har", url: "/api"})                    // HTTP Archive export
generate({format: "sarif"})                               // Accessibility SARIF report
generate({format: "csp", mode: "strict"})                 // Content Security Policy
generate({format: "sri"})                                 // Subresource Integrity hashes
generate({format: "pr_summary"})                          // PR performance summary

---

test — Playwright Regression Tests

Generates a complete Playwright test from the current browser session. Captures user actions, correlates them with API calls, and produces tests with real assertions — not just click replay.

generate({format: "test"})

generate({format: "test",
          test_name: "guest-checkout",
          base_url: "http://localhost:3000",
          assert_network: true,
          assert_no_errors: true,
          assert_response_shape: true})

Parameter	Type	Default	Description
test_name	string	Derived from URL	Name for the test() block
base_url	string	Captured origin	Replace origin in URLs for portability
assert_network	boolean	—	Include waitForResponse + status code assertions
assert_no_errors	boolean	—	Assert consoleErrors.length === 0
assert_response_shape	boolean	—	Assert response body structure matches (types only, never values)

What the output includes

• User actions translated to Playwright commands (click, fill, getByRole, etc.)
• Multi-strategy selectors prioritized: data-testid > ARIA role > label > text > ID > CSS
• Network assertions with waitForResponse and status code checks
• Response shape validation — field names and types, never actual values
• Console error collection — asserts zero errors during the flow
• Password redaction — passwords replaced with [user-provided]

Example output

import { test, expect } from '@playwright/test';

test('submit-form flow', async ({ page }) => {
  const consoleErrors = [];
  page.on('console', msg => {
    if (msg.type() === 'error') consoleErrors.push(msg.text());
  });

  await page.goto('http://localhost:3000/login');
  await page.getByLabel('Email').fill('user@example.com');
  await page.getByLabel('Password').fill('[user-provided]');

  const loginResp = page.waitForResponse(r => r.url().includes('/api/auth/login'));
  await page.getByRole('button', { name: 'Sign In' }).click();
  const resp = await loginResp;
  expect(resp.status()).toBe(200);

  expect(consoleErrors).toHaveLength(0);
});

---

reproduction — Bug Reproduction Scripts

Generates a Playwright script that reproduces the user’s actions leading up to a bug. Unlike test, this focuses on replaying the exact sequence rather than asserting outcomes.

generate({format: "reproduction"})

generate({format: "reproduction",
          error_message: "TypeError: Cannot read property 'id' of undefined",
          last_n: 10,
          base_url: "http://localhost:3000",
          include_screenshots: true})

Parameter	Type	Default	Description
error_message	string	—	Error message to include as context in the script
last_n	number	All	Use only the last N recorded actions
base_url	string	Captured origin	Replace origin in URLs
include_screenshots	boolean	—	Insert page.screenshot() calls between steps
generate_fixtures	boolean	—	Generate fixture files from captured network data
visual_assertions	boolean	—	Add toHaveScreenshot() assertions

---

har — HTTP Archive Export

Exports captured network traffic in HAR 1.2 format. HAR files can be imported into Chrome DevTools, Charles Proxy, or any HAR viewer for analysis.

generate({format: "har"})

generate({format: "har",
          url: "/api",
          method: "POST",
          status_min: 400,
          save_to: "/tmp/debug.har"})

Parameter	Type	Description
url	string	Filter by URL substring
method	string	Filter by HTTP method
status_min	number	Minimum status code
status_max	number	Maximum status code
save_to	string	File path to save the HAR file

---

sarif — Accessibility SARIF Report

Exports accessibility audit results in SARIF format (Static Analysis Results Interchange Format). SARIF files integrate with GitHub Code Scanning, VS Code, and CI/CD pipelines.

generate({format: "sarif"})

generate({format: "sarif",
          scope: "#main-content",
          include_passes: true,
          save_to: "/tmp/a11y.sarif"})

Parameter	Type	Description
scope	string	CSS selector to limit audit scope
include_passes	boolean	Include passing rules (not just violations)
save_to	string	File path to save the SARIF file

---

csp — Content Security Policy Generation

Generates a Content Security Policy header from observed network traffic. Gasoline sees which origins your page loads resources from and produces a CSP that allows exactly those origins.

generate({format: "csp"})

generate({format: "csp",
          mode: "strict",
          exclude_origins: ["https://analytics.google.com"],
          include_report_uri: true})

Parameter	Type	Description
mode	string	Strictness: strict, moderate, or report_only
exclude_origins	array	Origins to exclude from the generated CSP
include_report_uri	boolean	Include a report-uri directive

---

sri — Subresource Integrity Hashes

Generates SRI hashes for external scripts and stylesheets. SRI ensures that fetched resources haven’t been tampered with.

generate({format: "sri"})

generate({format: "sri",
          resource_types: ["script"],
          origins: ["https://cdn.example.com"]})

Parameter	Type	Description
resource_types	array	Filter: script, stylesheet
origins	array	Filter by specific CDN origins

---

pr_summary — PR Performance Summary

Generates a performance impact summary suitable for pull request descriptions. Compares before/after metrics and highlights regressions or improvements.

generate({format: "pr_summary"})

No additional parameters. Uses the current performance snapshot data.

Output includes:

• Web Vitals comparison (before/after)
• Regression/improvement verdicts
• Network request count changes
• Bundle size impact (if measurable)