01 · Roasts
Hackathon Hoarder
Six of your twelve analyzed repos are hackathon or single-sprint projects. You've got 6 hours of effort dressed up in ARCHITECTURE.md like it's a production system. The docs are longer than the code.
The Graveyard Gardener
staleRepoRatio = 0.34 and your heatmap is a ghost town before week 37. You were essentially offline for 8 months, then suddenly went feral. Seasonal developer or hibernating bear?
Stars? What Stars?
40 public repos, 496 commits this year, and a total star count of 8. You're building at full speed in a soundproof room. Even engine-visualizer, your empty scaffold with 0 KB and 0 commits, has more ambition than visibility.
Test-Optional Lifestyle
Out of 12 repos, exactly 3 have tests. You've got ARCHITECTURE.md, STATUS.md, design.md, AND docs/ in half your projects — but can't be bothered to write a single pytest. Documentation-driven development without the validation.
PyPI Ghost
locked_in_mcp is published to PyPI with an Apache-2.0 license, CI, and badges — all committed in 24 minutes. Either you're the fastest developer alive or that README is doing all the heavy lifting.
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% weight68C
- Consistency20% weight65C
- Quality20% weight69C
- Depth15% weight65C
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
98 active days
Language distribution
- Jupyter Notebook37%
- Python32%
- TypeScript22%
- JavaScript4%
- CSS1%
- HTML1%
- Other3%
04 · Numbers
Owned repos
non-fork
32
Commits
last 12 months
496
Followers
16
Joined GitHub
Feb 2022
05 · Top repos
BABTUNA /
Objaverse
A polished, end-to-end multimodal 3D search engine showcasing Daft's DataFrame pipeline pattern. Full-stack TypeScript+Python with semantic search, multi-view rendering, CLIP embeddings, LanceDB indexing, and interactive web frontend featuring search, UMAP atlas, and performance benchmarks.
BABTUNA /
yc_hud_hackathon
Hackathon RL environment (constrained text-only web-nav) with typed Python, comprehensive docs (ARCHITECTURE.md, STATUS.md, design.md), full test suite, and working agent training pipeline achieving 0.25→0.75 reward improvement on Qwen3-8B.
BABTUNA /
me_me_me
TypeScript Next.js 15 portfolio + blog with MDX, React Three Fiber 3D models, Tailwind CSS. Typed, documented, structured src/ layout, but no tests/CI. 138 KB codebase (~2.8k LOC). 30 commits in ~3 weeks shows burst work.
BABTUNA /
photon
Early-stage Rust/Python distributed runtime clone of Ray. Zero stars/forks, clearly experimental 3-week project with structured architecture, comprehensive docs, full test coverage via CI, but shipping only core abstractions without production use yet.
BABTUNA /
insforge_hackathon
AI-powered interior design assistant built in ~1 day (27 of last 30 commits in 6-hour window). 280KB TypeScript/Python codebase with documented architecture (ARCHITECTURE.md, STATUS.md) and clear Next.js+FastAPI stack, but no tests/CI, unpolished hackathon project with zero external adoption.
BABTUNA /
marvis
Early-stage React Three Fiber 3D visualization of meeting search ("Orrery galaxy"). TypeScript, Zustand state, TF-IDF search, optional Claude API integration. No tests, CI, or license; synthetic data generation included. Experimental prototype with refined UI but unproven adoption.
BABTUNA /
locked_in_mcp
LinkedIn MCP server using Playwright for browser automation; brand new (created 2026-04-15), minimal commits (~30), no source files sampled, but has README + docs (design.md, ARCHITECTURE.md) + tests + CI + license. Python project lacks type hints.
BABTUNA /
Applied-Intelligence-Hackathon
Hackathon Chrome extension project with clear architecture (Claude vision + InsForge logging) and comprehensive docs (design.md, ARCHITECTURE.md, STATUS.md). Demonstrates concept over production readiness; 15 commits in 4 hours, no tests/CI, untyped JavaScript, and no license.
BABTUNA /
Beta-Fund-X-Evermind-Hackathon-
Hackathon project: Chrome extension (EverNav) with vision-guided UI navigation via Claude & Evermind. Documented with design.md/ARCHITECTURE.md, 102 KB, built in ~4.5 hours. Zero production signals; no tests, CI, license, or typed code.
BABTUNA /
skim
Fresh Next.js+TypeScript code explorer / repository indexer. Typed, structured multi-file codebase with meaningful documentation but minimal visibility and only ~7 commits in <1 day. No tests, no CI, no license.
BABTUNA /
BABTUNA
Personal profile README with no actual code repository content. 971 KB appears to be profile visualization assets only; no source code files sampled. Zero production value.
BABTUNA /
engine-visualizer
Empty scaffold with zero commits, no documentation, no code, and no meaningful content. Repository created and immediately abandoned without any implementation.
06 · Timeline
- Feb 7, 2022Joined GitHub
- Aug 11, 2024Created BABTUNA
- 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
- May 13, 2026Created Objaverse
- May 22, 2026Created me_me_me — all about meeeeeeeeeeeeeeee
- May 24, 2026Created skim
- May 24, 2026Created photon
- May 30, 2026Created Beta-Fund-X-Evermind-Hackathon-
- May 31, 2026Created Applied-Intelligence-Hackathon
- Jun 3, 2026Created engine-visualizer
- Jun 6, 2026Created insforge_hackathon
- Jun 7, 2026Created marvis
- Jun 21, 2026Created yc_hud_hackathon
- Jun 24, 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.