s3dfx-cyber/gsoc-org-finder-

Purpose and Entry Points

GSoC 2026 Org Finder is a static, browser-first directory for discovering and comparing GSoC organizations by tags, domain, competition, activity, and beginner issues. The primary page is index.html; landing.html is a separate landing entry. Client bootstrapping and interaction wiring live in src/js/app.js, with presentation in src/styles.css.

Organization Catalog and Local Data

The application’s core domain model is the curated organization catalog, represented by src/js/org.js. It is supplemented by generated/static JSON artifacts for issue data, organization statistics, mentor data, and UI summaries:

• data/issues.json

• data/org-stats.json

• data/ui-summary.json

This boundary keeps primary discovery usable without live API availability while allowing data-refresh automation to update derived files.

Browser Application Layer

src/js/app.js owns the event-driven UI lifecycle: search, filtering, sorting, card rendering, modal behavior, compare selection, keyboard navigation, theme/preferences, and safe dynamic DOM rendering. It composes specialized feature modules:

• src/js/githubAnalyzer.js

• src/js/recommender.js

• src/js/recommendation-ui.js

User interactions transform catalog records plus local UI state into organization cards, detail dialogs, and a comparison view. Preferences and local analytics remain in browser localStorage; there is no account or central user database.

Matching and Recommendation Pipeline

Recommendation behavior is separated from base rendering. src/js/skillExtractor.js normalizes/extracts user skill context, while src/js/recommender.js scores organizations using skills, tags, competition, and experience-related signals. src/js/recommendation-ui.js presents the results. This makes the recommendation pipeline a client-only decision boundary: it consumes local catalog data and does not require the GitHub service.

GitHub Data Boundary

Live repository health and Good First Issue data flow from the browser through the Vercel Edge endpoint api/github.js, which proxies GitHub’s API and prevents exposing GITHUB_TOKEN to clients. It supports stats, Good First Issue counts, and bounded issue lists, with one-hour in-memory Edge caching. src/js/githubAnalyzer.js consumes this endpoint and enriches displayed organization data. GitHub availability/rate limits are isolated from baseline catalog browsing.

Offline and Deployment Runtime

The frontend is vanilla HTML/CSS/JavaScript with no framework, bundler, or application server. Vercel hosts static assets and runs the Edge Function, configured through vercel.json; .env.example documents required environment setup. sw.js provides offline resilience, using stale-while-revalidate for static assets and network-first behavior for API/JSON data. manifest.json supports installable/PWA-oriented browser metadata.

Data Maintenance and Governance Automation

Node scripts under agent/scripts/ maintain generated datasets, validate organization URLs and mentor data, and collect GitHub-derived inputs; notable stages include agent/scripts/fetch-issues.js, agent/scripts/generate-ui-cache.js, and agent/scripts/validate-generated-data.js. A separate Python TENET review agent resides in agent/tenet_agent/tenet_review.py. GitHub Actions in .github/workflows/ implement contribution intake, staged PR governance, mentor assignment/review tracking, and repository moderation; these are operational infrastructure rather than runtime application dependencies.