[BUG] Language breakdown fires up to 21 GitHub calls sequentially (N+1)

[vedant7007]

Describe the Bug

get_language_breakdown (backend/app/services/github_service.py:158) fetches the repo list, then loops up to 20 repos calling /repos/{full_name}/languages one at a time, awaiting each (:168-176) — up to 21 fully sequential round-trips per /github/languages call. The frontend hits this on every dashboard mount and every sync (DashboardPage.tsx:58). Result: multi-second loads and linear rate-limit burn with no caching.

Where

• Frontend (web dashboard)
• Backend (API)
• CLI (clutch-cli)

Steps to Reproduce

1. Connect a GitHub account with ~20 repos
2. Load the dashboard → triggers /github/languages
3. Observe the multi-second response time

Expected Behavior

Fire the per-repo language calls concurrently (single AsyncClient + asyncio.gather, bounded by a semaphore to stay polite to the rate limit), then aggregate — should be sub-second.

Screenshots / Terminal Output

n/a — measurable as endpoint latency scaling with repo count.

Environment

• OS: any
• Browser (if frontend): affects dashboard load
• CLI version (if CLI): n/a
• Commit: f3dd729

Additional Context

Could also add caching and guard each lang_response.json() for non-200. Happy to take it if you assign me.

Contributing as part of Elite Coders Summer of Code (ECSoC 2026).