01 · Roasts
Burst Coder, Not a Builder
6 of your top repos were built in 1–2 days. locked_in_mcp, marky, X_better, open_source_finder — all sprints. The heatmap looks like someone's heartbeat flatlined except for the last 10 weeks.
PyPI Without Stars
You shipped linkedin-scraper-mcp all the way to v4.9.1 on PyPI and still have 0 stars on the repo. Either you're very good at publishing or very bad at marketing — possibly both.
Tests Are a Myth
Only 1 of 9 scored repos has tests (locked_in_mcp). The other 8 are running on vibes and README confidence. 'HAS_TESTS=no' is your most consistent flag across the entire profile.
45 PRs, 9 Followers
You opened 45 external PRs this year — more than most engineers — yet only 9 people follow you. You're contributing everywhere and somehow staying completely invisible.
The Graveyard Ratio
45% of your repos were last pushed over 2 years ago. For someone who just started in 2022, that's impressive abandonment velocity. Each idea gets exactly one sprint before joining the cemetery.
Built using
Zoral
Shadows one worker for a week, then takes over their job with zero extra setup. Behaves exactly like the original.
zoral.ai
02 · Category breakdown
- Impact25% weight56D
- Consistency20% weight65C
- Quality20% weight72B
- Depth15% weight58D
- Breadth10% weight65C
- Community10% weight40D
03 · Stats
365-day commit heatmap
67 active days
Language distribution
- Jupyter Notebook51%
- Python36%
- TypeScript10%
- JavaScript2%
- Shell1%
- CSS1%
04 · Numbers
Owned repos
non-fork
22
Commits
last 12 months
178
Followers
9
Joined GitHub
Feb 2022
05 · Top repos
BABTUNA /
locked_in_mcp
LinkedIn MCP server enabling Claude to scrape profiles, companies, jobs via browser automation. Well-documented, tested, CI/CD configured, typed Python codebase with structured architecture and live PyPI package.
BABTUNA /
marky
Early-stage Python multi-agent orchestrator for ad research; integrates 5 intelligence agents (Local, Review, Yelp, Trends, RelatedQuestions) via Fetch.ai uAgents framework. Typed, well-documented with ARCHITECTURE.md, but nascent (1 day old, 10 of 30 commits), no tests/CI, no license.
BABTUNA /
founder_finder
Personal web scraper for Y Combinator founder profiles with CLI filtering and follow-assist tool. Single-file Python scripts using httpx + tqdm, clear docs, ~52 days active, 30 commits. Well-documented but untyped, no tests, no CI/license.
BABTUNA /
X_better
Personal Chrome extension for X/Twitter data collection (followers, tweets). Early-stage repo: 28 commits in 1 day, 386 KB untyped JS, documented README, no tests/CI. Functional feature-complete extension with GraphQL interception.
BABTUNA /
devfest-26
TypeScript hackathon project showcasing AI block marketplace with Flowglad billing integration, shipped with structured monorepo (frontend/backend/shared), docs (ARCHITECTURE.md, design.md, STATUS.md), and core features (blocks, checkout, webhooks, auth). Minimal adoption (1 star, 30 commits in 1 day), no tests/CI.
BABTUNA /
open_source_finder
Early-stage Python scraper pipeline for finding SF-based companies with open source projects. Typed, documented, structured multi-file layout with async GitHub scoring. Created 2026-04-11, 13 recent commits over ~1 day.
BABTUNA /
BABTUNA.github.io
Personal GitHub Pages site with minimal documentation. 30 commits over ~3.5 years with no README, tests, CI, or license. HTML-only static site showing modest sustained activity but limited scope.
BABTUNA /
yc_hackathon
One-shot hackathon dump created and pushed within 4 minutes. 5KB repo with README describing a paper scraper PoC, no tests, no CI, no license, Python without types. Experimental phase-1 project with no code samples visible.
BABTUNA /
BABTUNA
Personal profile README with no actual code repository content. 0 stars, 129 KB total size, CI workflow present but no source files, tests, or meaningful project implementation.
06 · Timeline
- Feb 7, 2022Joined GitHub
- Oct 27, 2022Created BABTUNA.github.io
- Aug 11, 2024Created BABTUNA
- Jan 31, 2026Created marky
- Feb 7, 2026Created devfest-26
- Feb 22, 2026Created founder_finder
- Mar 1, 2026Created yc_hackathon — gooning rn
- Mar 28, 2026Created X_better — X lowkey kinda sucks. Let's fix that.
- Apr 11, 2026Created open_source_finder — finding good open sources to work on is lowkey hella tuff
- Apr 15, 2026Created locked_in_mcp — ayo lowkey trying to use playwright to scrape the big Link was ass so we went with Claude MCP. TS is straight gas ngl
- Apr 28, 2026Most recent push to BABTUNA
07 · Compare
08 · Rubric
How this score was produced
Overall = Σ (category × weight) + gentle top-end curve
Tier thresholds
▸ How the pipeline works
- 01Scrape.Pull every non-fork repo pushed in the last 90 days, plus your contribution calendar, followers, and language byte counts — straight from GitHub's REST & GraphQL APIs.
- 02Triage.A small model reads every repo's file tree + README and picks the 20 files per repo that actually reveal how you code.
- 03Grade each repo. All repos run in parallel through a fast scoring model that reads the picked files and rates each one independently on Impact, Quality, and Depth — with evidence citations.
- 04Aggregate. A larger reasoning model combines the per-repo scores with server-computed stats (heatmap, commit cadence, language entropy, follower count) to produce the 6-dimension profile score + roasts.
- 05Correct.Deterministic server-side checks enforce anchor-scale floors (e.g. a profile with 2,000+ public commits can't score 30 Consistency) and recompute the final verdict.
~90 seconds per profile, ~$0.25 in compute. Total of ~240 files read across your top-12 repos. One rating per GitHub account per day.
▸ Data sources & caveats
- Heatmap & commit totals: GitHub GraphQL
contributionsCollection— covers the last 365 days, includes private repos when the user has opted in (default). - Language %: byte totals across the top 30 owned non-fork repos.
- Curve: a small upward nudge centered on raw score ≈ 70, capping at 100. Prevents specialists from being unfairly penalised for narrow breadth.
- Anchor corrections: when server-measured signals (e.g. privateWorkLikely, multiRepoVolume, follower count) mandate a minimum category score, the aggregation step enforces it. These are signal-conditional, not identity-based floors.