VicroCode
Make Code Create Value
VicroCode is a lightweight online platform for publishing, running, sharing, and monetizing code projects. Launch HTML, Python, SQLite, AI agents, management tools, games, and more without server setup.
Please wait while VicroCode loads. You can also explore the AI programming guide
Loading...

AI MARKET GUIDE

Rebuild a GitHub Actions Search Tool as a Hosted HTML Workflow Browser

Strip the Chrome extension wrapper from a workflow search-and-favorites pattern and rebuild it as a shareable HTML tool with SQLite persistence.

A Chrome extension built for GitHub Actions workflow search and favorites demonstrates a clear pattern: users need to filter, flag, and return to specific workflows without scrolling through long lists. The extension solves that by injecting search and star controls into the GitHub sidebar. But the useful core—search logic, favorites storage, and retrieval—doesn't depend on Chrome APIs or live GitHub DOM manipulation. You can extract that core, rebuild it as a standalone HTML tool, and host it where anyone can use it without installing an extension.

One developer shipped Workflow Harbor, a Chrome extension that adds search and repository-scoped favorites to GitHub Actions workflow lists. The extension handles sidebar pagination, input method composition delays, and cross-device sync through Chrome's storage API. The reported workflow shows the mechanics work: name and filename filtering, persistent star flags per repository, and UI that adapts to GitHub's sidebar states. But it also highlights the boundary between what requires a browser extension and what doesn't. The search algorithm, the favorites data structure, and the display logic are portable. The Chrome manifest, content script injection, and sync API are not.

What can you rebuild without the extension layer?

You can rebuild the search-and-favorites workflow as a hosted HTML application that accepts workflow file uploads, filters by name or filename, persists favorites in SQLite, and renders results in a static interface. Use AI coding to structure the implementation: an HTML interface with file input, a JavaScript module that parses workflow YAML or JSON, a search function that matches names and filenames, and a SQLite table that stores repository slugs, workflow identifiers, and star timestamps. The tool lets users run HTML online without backend deployment, and you can edit the favorites table directly with the SQLite editor during development or debugging.

The trade-off is immediacy. The extension reads workflows live from the GitHub DOM; your hosted tool requires users to export or copy workflow files first. But the hosted version runs anywhere, doesn't require installation permissions, and gives you full control over the data model. If you store workflow metadata—file path, last modified date, commit hash—you can build a changelog view or a workflow history diff without touching GitHub's API rate limits.

Why would someone use a standalone workflow browser instead of the extension?

A standalone tool decouples the workflow search pattern from GitHub's sidebar implementation. The extension must track GitHub's DOM structure, handle pagination edge cases, and adapt to future sidebar redesigns. A hosted HTML tool just reads uploaded files, applies filters, and displays results. If GitHub changes the sidebar again, the hosted version keeps working. If you need to search workflows from multiple repositories at once, you can batch-upload files and filter across the entire set. If you want to share a curated list of workflows with a team that doesn't use Chrome, you export the SQLite database and send a link.

The hosted tool also exposes the favorites data as a queryable table. You can write SQL to count stars per repository, find workflows you haven't opened in six months, or generate a report of the most-starred workflows across projects. The extension stores favorites in Chrome's sync storage, which syncs across devices but doesn't expose a query interface. SQLite gives you schema control, indexing, and joins.

How do you map workflow file structure to search and favorites?

GitHub Actions workflows use YAML with a predictable schema: a `name` field, a `on` trigger block, and a `jobs` map. Your HTML tool parses the uploaded YAML, extracts the `name` and the filename, and stores both in a JavaScript object. The search function checks whether the user's query substring matches either field, case-insensitive. The favorites table stores `repository`, `workflow_id`, and `starred_at`. When the user clicks a star icon, you insert a row; when they unstar, you delete it. When rendering the workflow list, you join the parsed workflows with the favorites table to display star status.

One friction point: workflow identity. The extension can use GitHub's workflow ID from the DOM. Your hosted tool doesn't have that, so you define identity yourself. A composite key of repository slug and workflow filename works if filenames are unique per repository. If users rename workflows, you'll need to decide whether to treat the renamed file as a new workflow or preserve the old star. The simplest approach is filename-based identity and explicit re-starring after renames. Document that trade-off in the tool's help text.

What does AI coding handle in this rebuild?

Use AI coding to generate the YAML parser, the search filter function, and the SQLite schema. Describe the workflow structure, paste a sample workflow file, and ask for a parsing function that extracts `name`, `on`, `jobs`, and filename. The AI writes the JavaScript, handles edge cases like missing `name` fields or multiline strings, and suggests a fallback display name when `name` is absent. For the SQLite schema, describe the favorites model and ask for a `CREATE TABLE` statement with indexes on `repository` and `workflow_id`. The AI generates the schema, suggests an additional index on `starred_at` for chronological queries, and writes the `INSERT` and `DELETE` statements for star toggles.

The HTML interface benefits from AI-generated boilerplate: file input handling, drag-and-drop upload, and result rendering. Describe the layout—file upload zone at the top, search bar below, results list with star icons—and the AI writes the HTML structure, CSS for layout and star icon styling, and JavaScript event listeners for upload, search input, and star clicks. You review the generated code, adjust styling, and test with a batch of real workflow files. If the search feels sluggish with hundreds of workflows, ask the AI to add debouncing to the search input handler. If the star icon doesn't visually toggle, describe the expected behavior and the AI updates the CSS or adds a class toggle.

What stays out of scope?

Live GitHub integration, Chrome extension behavior, and cross-device sync through browser APIs are explicitly out of scope. VicroCode supports HTML execution and SQLite persistence, but it doesn't provide GitHub OAuth, REST API wrappers, or Chrome extension manifests. If you need to pull workflows directly from GitHub repositories, you'll add GitHub's API client separately and handle authentication outside the hosted tool. If you want sync, you export the SQLite database and re-import it on another device, or you build a separate sync mechanism using an API endpoint hosted on VicroCode. The platform handles API endpoint hosting and in-platform tool calls, so you could expose a `/favorites` endpoint that accepts JSON and writes to SQLite, then call that endpoint from multiple tool instances. But that's a second step, not part of the initial rebuild.

The hosted tool also doesn't handle pagination the way the extension does. GitHub's workflow list paginates server-side; your tool displays all uploaded workflows in one scrollable list. If someone uploads 500 workflows, the list renders all 500. You can add client-side pagination or virtual scrolling if performance degrades, but that's a UI optimization, not a core feature. The search pattern already filters the list, so users rarely see all 500 at once.

FAQ

Can I monetize a hosted workflow search tool on VicroCode?

Yes, VicroCode supports project publishing, hosting, sharing, and monetization. You can charge for access, offer a free tier with limited uploads, or monetize through usage-based pricing tied to the number of repositories or workflows.

Does the SQLite database persist across sessions?

Yes, VicroCode's SQLite databases persist per project. Your favorites remain available between visits, and you can export the database file for backup or migration.

What happens if I upload a workflow file with no name field?

Your parser should fall back to the filename when the `name` field is missing. Document that behavior in the tool's help text so users understand how unnamed workflows display.

Further reading