01 · Roasts
Sprint, then vanish
kawa has 19 sampled commits, while nix-vps arrived as a one-commit deployment manifesto.
CI is the missing service
All three highlighted repos lack CI; even gambit-rs has integration tests waiting for a runner.
Architecture enjoys cardio
Kawa has four Compose services and a roadmap; its test suite remains a TODO.
Protocol archaeologist
gambit-rs documents unsupported wire shapes and redacts secrets, then stops after 3 sampled commits.
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% weight36F
- Consistency20% weight35F
- Quality20% weight57D
- Depth15% weight35F
- Breadth10% weight65C
- Community10% weight50D
03 · Stats
365-day commit heatmap
305 active days
Language distribution
- Jupyter Notebook24%
- C++19%
- TypeScript16%
- JavaScript10%
- Rust10%
- Go7%
- Other14%
04 · Numbers
Owned repos
non-fork
25
Commits
last 12 months
56
Followers
50
Joined GitHub
Feb 2017
05 · Top repos
sanstzu /
gambit-rs
A well-documented, typed Rust async client with guarded WebSocket gameplay, protocol notes, and local integration tests, but it is a one-day, 1-star project with no demonstrated external adoption.
sanstzu /
kawa
A documented Rust livestream platform split into ingest, transcoder, and session-manager services, with Docker Compose deployment and RTMP/gRPC/Redis/S3 integration, but limited visible adoption and no tests, CI, or license.
sanstzu /
nix-vps
A well-documented NixOS VPS deployment scaffold with reusable container, nginx proxy, disko, and sops-nix modules, but currently a zero-star one-commit project without tests, CI, or licensing.
06 · Timeline
- Feb 28, 2017Joined GitHub
- Jan 18, 2024Created kawa — A simple livestream application fully written in Rust
- Jul 3, 2026Created nix-vps
- Aug 25, 2026Created gambit-rs — Unofficial Rust Client for gambit.com
- Aug 25, 2026Most recent push to gambit-rs
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.