docs(skills): refine snapshot diff and QuickJS guidance (#350)

* benchmakrs added

* fix: removed missing refs from readme

* feat(benchmarks): support OpenAI as judge provider

* fix: screenshot upload

* feat: added cost support

* docs: design benchmark-specific judge prompts

* feat(benchmarks): split general and stealth judge prompts

* Delete docs/superpowers/specs directory

* chore: bump version metadata for 0.7.0 release

* Revert "chore: bump version metadata for 0.7.0 release"

This reverts commit c69011f.

* feat(benchmarks): expose full Webcmd skill pack to Codex evals

* feat(benchmarks): clarify stealth captcha handling

* fix: added support for webcmd browser skills only

* feat(benchmarks): run eval attempts in parallel

* fix(benchmarks): isolate parallel webcmd sessions

* test(benchmarks): cover parallel webcmd sessions

* pi support (#313)

* chore(benchmarks): pin webcmd 0.7.1

* refactor: disable concurrency in benchmark evaluation and remove integration tests

* feat(benchmarks): support Codex subscription evals

* docs(skills): improve browser diff and form guidance

* docs(skills): sync browser skill sources

* docs(skills): refine snapshot diff and QuickJS guidance

---------

Co-authored-by: beubax <tejasr@bu.edu>
Co-authored-by: Tejas <48682479+beubax@users.noreply.github.com>

Author: rishabhraj36

skill-src/webcmd-browser/SKILL.src.md

@@ -184,13 +184,13 @@ For CAPTCHA or raw user takeover, stop automation, give the user any viewer URL
 
 For native controls, inspect structure first, then use Playwright's normal form APIs inside `run`. Do not guess date formats, option labels, or file constraints from memory; read them from the DOM.
 
-`browser run` uses QuickJS, not Node.js: `Buffer` and `TextEncoder` are unavailable. For an in-memory upload, pass bytes directly:
+`browser run` uses QuickJS, not Node.js: `Buffer` is unavailable. For an in-memory upload, pass a `Uint8Array`; encode text with the supported `TextEncoder`:
 
 ```js
 await page.locator('input[type="file"]').setInputFiles({
   name: 'evidence.txt',
   mimeType: 'text/plain',
-  buffer: new Uint8Array([102, 105, 108, 101]),
+  buffer: new TextEncoder().encode('file'),
 });
 ```
 

skills/webcmd-browser/SKILL.md

@@ -184,13 +184,13 @@ For CAPTCHA or raw user takeover, stop automation, give the user any viewer URL
 
 For native controls, inspect structure first, then use Playwright's normal form APIs inside `run`. Do not guess date formats, option labels, or file constraints from memory; read them from the DOM.
 
-`browser run` uses QuickJS, not Node.js: `Buffer` and `TextEncoder` are unavailable. For an in-memory upload, pass bytes directly:
+`browser run` uses QuickJS, not Node.js: `Buffer` is unavailable. For an in-memory upload, pass a `Uint8Array`; encode text with the supported `TextEncoder`:
 
 ```js
 await page.locator('input[type="file"]').setInputFiles({
   name: 'evidence.txt',
   mimeType: 'text/plain',
-  buffer: new Uint8Array([102, 105, 108, 101]),
+  buffer: new TextEncoder().encode('file'),
 });
 ```
 

src/browser/run/runner.test.ts

@@ -20,6 +20,7 @@ import {
 } from 'playwright-core';
 import { LocalBrowserRunArtifactSink } from './artifacts.js';
 import { MemorySnapshotBaselineStore } from '../snapshot/index.js';
+import { MemorySnapshotBaselineStore } from '../snapshot/index.js';
 import { unsupportedApiMessage } from './playwright-transport.js';
 import { QuickJSHost } from './quickjs-host.js';
 import { runBrowserProgram } from './runner.js';