Skip to content

[DASH-1732] [templates] add build step to create javascript templates#75

Open
syd-shields wants to merge 24 commits into
devfrom
sydshields/dash-1732-turn-existing-templates-in-js-files
Open

[DASH-1732] [templates] add build step to create javascript templates#75
syd-shields wants to merge 24 commits into
devfrom
sydshields/dash-1732-turn-existing-templates-in-js-files

Conversation

@syd-shields
Copy link
Copy Markdown

@syd-shields syd-shields commented Mar 30, 2026

This PR adds a playground-production CI workflow that builds Typescript templates into playground compatible templates as deemed by the website-api templates API.

In order to marke this Linear issue as done:

  1. merge sydshields/dash-1732-turn-existing-templates-in-js-files into dev
  2. create a new production branch, enable branch protection
  3. merge dev into production

Added a JS build script

  • Added a build script in package.json build:javascript. This command runs build TS files into JS ones in CI
  • Added scripts/build-javascript.mjs which builds the JS files to /javascript

Added a production branch for dashboard use

  • Added GitHub Actions workflow at .github/workflows/playground-productiont.yml. This only runs on the production branch
  • Added scripts/playground-ci.mjs which is ran in the playground-production workflow. It calls fetchPlaygroundTypescriptDirNames and runs scripts/build-javascript.mjs

Ensure playground compatibility by delegating templates API as source of truth

  • Added a function fetchPlaygroundTypescriptDirNames which uses the wesbite-api templates API to grab playground compatible templates from the source of truth
  • Add hasStagehandUsage to ensure Stagehand is being used correctly

What's not included:

  • Parsing templates so they can actually run in the playground (removing unused vars, BB/Stagehand configs, this will be handled dash side, since it has more to do with the environment the code will run on (sandbox IDE).

Note

Medium Risk
Adds new GitHub Actions workflows that fetch a remote templates list, generate javascript/ output, and auto-commit/push to protected branches; failures or API changes could impact release CI and repo state.

Overview
Introduces a playground-only TypeScript→JavaScript generation pipeline: pnpm run ci:playground now builds javascript/ from a filtered set of typescript/ templates determined by the public templates API, then validates those sources.

Adds reusable GitHub Actions (.github/workflows/playground.yml) plus branch-specific triggers for production and test-production that run the pipeline and auto-commit/push regenerated javascript/ on pushes.

Adds local tooling via scripts/build-javascript.mjs (pnpm run build:javascript) with include/exclude filters and package.json rewriting, plus validation checks in scripts/validate-playground-templates.mjs (parse diagnostics + forbidden connectOverCDP usage + Stagehand constructor enforcement). Also adds CODEOWNERS coverage for the new workflows/scripts and documents the new source-of-truth behavior in README.md.

Reviewed by Cursor Bugbot for commit 3e85383. Bugbot is set up for automated code reviews on this repo. Configure here.

@syd-shields
Copy link
Copy Markdown
Author

syd-shields commented Mar 30, 2026

This change is part of the following stack:

Change managed by git-spice.

@syd-shields syd-shields changed the title add build step to create javascript templates [DASH-1732] [templates] add build step to create javascript templates Mar 30, 2026
@syd-shields syd-shields requested a review from jay-sahnan April 1, 2026 14:38
Comment thread scripts/build-javascript.mjs
@jay-sahnan
Copy link
Copy Markdown
Contributor

In the templates repo, all scripts should be runnable out of the box.

Some templates have package.json files that need to be copied over to the JavaScript versions. However, we can't just copy them as-is — they still reference TypeScript. For example, they contain "start": "tsx index.ts" and TypeScript devDependencies. These need to be rewritten for JS (e.g., "start": "node index.js", remove tsx, typescript, @types/*).

We should also copy over the READMEs if possible.

Comment thread scripts/build-javascript.mjs
Comment thread scripts/build-javascript.mjs Outdated
Copy link
Copy Markdown

@cursor cursor Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Autofix Details

Bugbot Autofix prepared a fix for the issue found in the latest run.

  • ✅ Fixed: Build script copies READMEs without transforming TypeScript references
    • The build script now rewrites README content when copying so TypeScript paths/extensions are converted to JavaScript equivalents and regenerated JavaScript READMEs contain correct javascript/ and .js references.

Create PR

Or push these changes by commenting:

@cursor push 94170f2cf4
Preview (94170f2cf4)
diff --git a/javascript/amazon-product-scraping/README.md b/javascript/amazon-product-scraping/README.md
--- a/javascript/amazon-product-scraping/README.md
+++ b/javascript/amazon-product-scraping/README.md
@@ -17,11 +17,11 @@
 
 ## QUICKSTART
 
-1. cd typescript/amazon-product-scraping
+1. cd javascript/amazon-product-scraping
 2. npm install
 3. cp .env.example .env (or create .env with required keys)
 4. Add BROWSERBASE_PROJECT_ID, BROWSERBASE_API_KEY, and GOOGLE_API_KEY to .env
-5. Optionally edit SEARCH_QUERY in index.ts
+5. Optionally edit SEARCH_QUERY in index.js
 6. npm start
 
 ## EXPECTED OUTPUT
@@ -49,7 +49,7 @@
 
 ## NEXT STEPS
 
-• Switch to direct URL: Uncomment the URL-based search block in index.ts for faster runs without LLM search actions.
+• Switch to direct URL: Uncomment the URL-based search block in index.js for faster runs without LLM search actions.
 • Parameterize query: Accept SEARCH_QUERY from CLI or env for different products without editing code.
 • Paginate: Extend extraction to multiple pages or increase the number of products per run.
 

diff --git a/javascript/company-address-finder/README.md b/javascript/company-address-finder/README.md
--- a/javascript/company-address-finder/README.md
+++ b/javascript/company-address-finder/README.md
@@ -27,7 +27,7 @@
 2. npm install
 3. cp .env.example .env
 4. Add your Browserbase API key, Project ID, and Google Generative AI API key to .env
-5. Edit COMPANY_NAMES array in index.ts to specify which companies to process
+5. Edit COMPANY_NAMES array in index.js to specify which companies to process
 6. npm start
 
 ## EXPECTED OUTPUT

diff --git a/javascript/company-value-prop-generator/README.md b/javascript/company-value-prop-generator/README.md
--- a/javascript/company-value-prop-generator/README.md
+++ b/javascript/company-value-prop-generator/README.md
@@ -16,7 +16,7 @@
 
 ## QUICKSTART
 
-1. cd typescript/company-value-prop-generator
+1. cd javascript/company-value-prop-generator
 2. npm install
 3. cp .env.example .env
 4. Add required API keys/IDs to .env

diff --git a/javascript/dynamic-form-filling/README.md b/javascript/dynamic-form-filling/README.md
--- a/javascript/dynamic-form-filling/README.md
+++ b/javascript/dynamic-form-filling/README.md
@@ -21,7 +21,7 @@
 1. pnpm install
 2. cp .env.example .env
 3. Add required API keys/IDs to .env (BROWSERBASE_PROJECT_ID, BROWSERBASE_API_KEY, GOOGLE_GENERATIVE_AI_API_KEY)
-4. Customize the `tripDetails` variable in index.ts with your own form data
+4. Customize the `tripDetails` variable in index.js with your own form data
 5. Update the form URL if using a different form
 6. pnpm start
 

diff --git a/javascript/extend-browserbase/README.md b/javascript/extend-browserbase/README.md
--- a/javascript/extend-browserbase/README.md
+++ b/javascript/extend-browserbase/README.md
@@ -23,7 +23,7 @@
 
 ## QUICKSTART
 
-1. cd typescript/extend-browserbase
+1. cd javascript/extend-browserbase
 2. pnpm install
 3. cp .env.example .env
 4. Add required API keys to .env:

diff --git a/javascript/getting-started-with-browserbase/README.md b/javascript/getting-started-with-browserbase/README.md
--- a/javascript/getting-started-with-browserbase/README.md
+++ b/javascript/getting-started-with-browserbase/README.md
@@ -31,7 +31,7 @@
 
 ### TypeScript
 
-1. cd typescript/getting-started-with-browserbase
+1. cd javascript/getting-started-with-browserbase
 2. npm install
 3. cp .env.example .env
 4. Add BROWSERBASE_API_KEY to .env (get it from https://browserbase.com/settings)

diff --git a/javascript/image-url-download/README.md b/javascript/image-url-download/README.md
--- a/javascript/image-url-download/README.md
+++ b/javascript/image-url-download/README.md
@@ -15,7 +15,7 @@
   Docs → https://docs.stagehand.dev/basics/extract
 - page.evaluate: run a JavaScript function directly inside the browser context — inherits proxy, cookies, and headers.
   Docs → https://playwright.dev/docs/evaluating
-- MAX_IMAGES: configurable cap (default: 10) on how many images to download per run. Set via the `MAX_IMAGES` env var or the constant at the top of `index.ts`.
+- MAX_IMAGES: configurable cap (default: 10) on how many images to download per run. Set via the `MAX_IMAGES` env var or the constant at the top of `index.js`.
 
 ## QUICKSTART
 
@@ -43,7 +43,7 @@
 - Empty images folder: some pages load images lazily — try scrolling the page before extraction, or increase the page load wait
 - Zero images found: the page may use CSS background images not captured by `<img>` tags — adjust the extract instruction to target specific selectors
 - CORS / auth-gated images: images behind login walls or strict CORS policies may fail in `page.evaluate()` — ensure you are authenticated before running the script
-- MAX_IMAGES cap: if you need more than 10 images, set `MAX_IMAGES=50` in your .env or edit the constant at the top of `index.ts`
+- MAX_IMAGES cap: if you need more than 10 images, set `MAX_IMAGES=50` in your .env or edit the constant at the top of `index.js`
 - Large pages: pages with hundreds of images may slow down `extract()` — use MAX_IMAGES to limit the download set
 
 ## USE CASES
@@ -58,7 +58,7 @@
 • Scroll before extracting: call `page.evaluate(() => window.scrollTo(0, document.body.scrollHeight))` before `extract()` to trigger lazy-loaded images.
 • Concurrent downloads: fan out the `page.evaluate` fetch calls with `Promise.allSettled` for faster bulk downloads.
 • Metadata CSV: write a `manifest.csv` alongside the images recording original URL, filename, MIME type, byte size, and download timestamp.
-• Extend MIME support: add entries to the `MIME_TO_EXT` map at the top of `index.ts` for any formats not already covered.
+• Extend MIME support: add entries to the `MIME_TO_EXT` map at the top of `index.js` for any formats not already covered.
 
 ## HELPFUL RESOURCES
 

diff --git a/javascript/playwright/basic-recaptcha/README.md b/javascript/playwright/basic-recaptcha/README.md
--- a/javascript/playwright/basic-recaptcha/README.md
+++ b/javascript/playwright/basic-recaptcha/README.md
@@ -67,7 +67,7 @@
 
 ## QUICKSTART
 
-1. cd typescript/playwright/basic-recaptcha
+1. cd javascript/playwright/basic-recaptcha
 2. pnpm install
 3. cp .env.example .env
 4. Add your Browserbase API key to .env

diff --git a/javascript/sec-filing-research/README.md b/javascript/sec-filing-research/README.md
--- a/javascript/sec-filing-research/README.md
+++ b/javascript/sec-filing-research/README.md
@@ -25,7 +25,7 @@
 2. npm install
 3. cp .env.example .env
 4. Add BROWSERBASE_PROJECT_ID, BROWSERBASE_API_KEY, and GOOGLE_API_KEY to .env
-5. (Optional) Edit SEARCH_QUERY and NUM_FILINGS in index.ts
+5. (Optional) Edit SEARCH_QUERY and NUM_FILINGS in index.js
 6. npm start
 
 ## EXPECTED OUTPUT

diff --git a/javascript/website-link-tester/README.md b/javascript/website-link-tester/README.md
--- a/javascript/website-link-tester/README.md
+++ b/javascript/website-link-tester/README.md
@@ -18,7 +18,7 @@
 ### QUICKSTART
 
 1. **cd into the template**
-   - `cd typescript/website-link-tester`
+   - `cd javascript/website-link-tester`
 2. **Install dependencies**
    - `npm install`
 3. **Configure environment**
@@ -76,7 +76,7 @@
 
 ### TUNING BATCH SIZE & CONCURRENCY
 
-- **`MAX_CONCURRENT_LINKS` in `index.ts`**
+- **`MAX_CONCURRENT_LINKS` in `index.js`**
   - Default: `1` → sequential link verification (works on all plans)
   - Set to `> 1` → more concurrent link verifications per batch (requires higher Browserbase concurrency limits)
 - **Using Semaphores for advanced control**

diff --git a/scripts/build-javascript.mjs b/scripts/build-javascript.mjs
--- a/scripts/build-javascript.mjs
+++ b/scripts/build-javascript.mjs
@@ -76,6 +76,15 @@
   return pkg;
 }
 
+function adaptReadmeForJavaScript(content) {
+  return content
+    .replaceAll("typescript/", "javascript/")
+    .replaceAll(/\.d\.ts\b/g, "__PRESERVE_D_TS__")
+    .replaceAll(/\.tsx\b/g, ".jsx")
+    .replaceAll(/\.ts\b/g, ".js")
+    .replaceAll(/__PRESERVE_D_TS__/g, ".d.ts");
+}
+
 function normalizeRelativePath(filePath) {
   return filePath.split(path.sep).join("/");
 }
@@ -187,6 +196,15 @@
   await writeFile(outputPath, `${JSON.stringify(adapted, null, 2)}\n`, "utf8");
 }
 
+async function writeAdaptedReadme(sourcePath) {
+  const relativePath = path.relative(TYPESCRIPT_DIR, sourcePath);
+  const outputPath = path.join(JAVASCRIPT_DIR, relativePath);
+  const source = await readFile(sourcePath, "utf8");
+  const adapted = adaptReadmeForJavaScript(source);
+  await mkdir(path.dirname(outputPath), { recursive: true });
+  await writeFile(outputPath, adapted, "utf8");
+}
+
 async function transpileFile(sourcePath) {
   const relativePath = path.relative(TYPESCRIPT_DIR, sourcePath);
   const outputPath = path
@@ -260,6 +278,8 @@
     assetFiles.map(async (filePath) => {
       if (path.basename(filePath) === "package.json") {
         await writeAdaptedPackageJson(filePath);
+      } else if (path.basename(filePath) === "README.md") {
+        await writeAdaptedReadme(filePath);
       } else {
         await copyAssetFile(filePath);
       }

This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.

Comment thread scripts/build-javascript.mjs
@syd-shields syd-shields force-pushed the sydshields/dash-1732-turn-existing-templates-in-js-files branch from 8d8479e to 3173ce4 Compare May 4, 2026 15:26
Comment thread scripts/build-javascript.mjs
Comment thread scripts/build-javascript.mjs
Comment thread scripts/playground-ci.mjs Outdated
const argv = ["scripts/build-javascript.mjs"];
for (const name of dirs) {
argv.push("--include-template", name);
}
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Empty API response causes CI to build all templates

Medium Severity

When fetchPlaygroundTypescriptDirNames returns an empty array (e.g., the templates API returns zero playground-runnable templates), the for loop adds no --include-template flags to argv. The build script then receives no filter arguments, causing includeTemplates.size to be 0 and falling through to the else branch that wipes javascript/ and rebuilds every template instead of none. The CI workflow would then commit all templates to the production branch, which is the opposite of the intended behavior.

Additional Locations (1)
Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit b17759b. Configure here.

Comment thread scripts/playground-ci.mjs Outdated
const argv = ["scripts/build-javascript.mjs"];
for (const name of dirs) {
argv.push("--include-template", name);
}
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Empty API response causes CI to build all templates

Medium Severity

When fetchPlaygroundTypescriptDirNames returns an empty array (e.g., the templates API returns zero playground-runnable templates), the for loop adds no --include-template flags to argv. The build script then receives no filter arguments, causing includeTemplates.size to be 0 and falling through to the else branch that wipes javascript/ and rebuilds every template instead of none. The CI workflow would then commit all templates to the production branch, which is the opposite of the intended behavior.

Additional Locations (1)
Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit b17759b. Configure here.


const includeTemplates = new Set(parseListFlag(argv, "--include-template"));
const includePaths = parseListFlag(argv, "--include-path");
const excludePathOnly = parseListFlag(argv, "--exclude-path");
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Flags parsed redundantly in filter and main function

Low Severity

parseListFlag is called for --include-template, --include-path, and --exclude-path in buildJavaScriptTemplates (lines 29–31), duplicating the exact same parsing already done inside createFileFilter (lines 92, 98, 99). This duplication means changes to flag names or parsing logic need to be synchronized in two places — if one is updated and the other is not, the file-filtering and directory-cleanup behaviors would silently diverge.

Additional Locations (1)
Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit ed33a0c. Configure here.

}
if (typeof scripts.build === "string" && /^\s*tsc(\s|$)/u.test(scripts.build)) {
delete scripts.build;
}
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Build script deletion too aggressive for compound commands

Medium Severity

The regex /^\s*tsc(\s|$)/u deletes the entire build script if it merely starts with tsc, even in compound commands like "tsc && next build" or "tsc --noEmit && vite build". The (\s|$) after tsc matches a space, so tsc && next build passes the test and the whole script is removed — silently dropping the non-TypeScript next build step. The check intends to strip pure-tsc build scripts but doesn't account for &&/;/|| chained commands.

Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit ed33a0c. Configure here.

Comment thread javascript/google-trends/package.json Outdated
Comment thread .github/workflows/playground-test-production.yml Outdated
Copy link
Copy Markdown

@cursor cursor Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cursor Bugbot has reviewed your changes and found 1 potential issue.

There are 5 total unresolved issues (including 4 from previous reviews).

Fix All in Cursor

❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.

Reviewed by Cursor Bugbot for commit a143d03. Configure here.

Comment thread .github/workflows/playground.yml Outdated
Copy link
Copy Markdown
Contributor

@jay-sahnan jay-sahnan left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Syd, this looks great! Thank you!

Tiny NITS:

  1. We should add javascript/ to .gitignore
  2. How do we test/validate that the conversion to JS? Is this a CI job on future PRs to the dev branch?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants