01 · Roasts
The One-Day Magician
5 of your 7 repos were created AND last-pushed on 2026-03-21. You apparently deposited an entire portfolio in a single afternoon. GitHub history is not a dump truck, Randev.
Burst Fire, Long Silence
Your heatmap is dead for 35 of 52 weeks, then suddenly erupts with wall-to-wall 4s for a few weeks. That's not consistency — that's a developer who remembers GitHub exists once a semester.
346 Tests, Zero Friends
Automaton-QTS ships with 346 unit tests, strict mypy, and a full backtest engine — and has 1 star (probably yours). All that rigor, and the world hasn't noticed yet.
180 PRs, 11 Followers
You filed 180 pull requests this year but only convinced 11 people to follow you. That's either very targeted corporate contribution or the most under-marketed shipping record on this platform.
School Project Apologist
Poem-Master's README literally says 'school project for promoting reading.' It's 3KB, has one commit, and uses console.log as error handling. At least the Heroku backend has moved on.
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% weight55D
- Consistency20% weight35F
- Quality20% weight62C
- Depth15% weight50D
- Breadth10% weight65C
- Community10% weight40D
03 · Stats
365-day commit heatmap
44 active days
Language distribution
- Python49%
- C++33%
- Shell7%
- TypeScript5%
- JavaScript3%
- CMake1%
- Other2%
04 · Numbers
Owned repos
non-fork
11
Commits
last 12 months
356
Followers
11
Joined GitHub
Feb 2018
05 · Top repos
SparkleButt747 /
Claudius-Maximus
Experimental benchmark optimization harness for Claude Code on Harbor benchmarks, achieving 100% pass rate on Terminal-Bench 2.0 through structured prompts, hook-based monitoring, and middleware pattern injection—novel work but early-stage with minimal adoption.
SparkleButt747 /
Automaton-QTS
Early-stage trading engine (334 KB) with strong architecture—typed Python, comprehensive signals (RSI/MACD/HMM/sentiment), human-in-loop LLM oversight, and risk controls—but only 1 star, minimal external adoption, and recent creation (11 days old). Shipping as-is demonstrates capability.
SparkleButt747 /
RAG
Personal RAG system with multi-format ingestion, hybrid retrieval, neural reranking, and policy gating. Typed Python (3.12), structured architecture (chunkers, extractors, agents), documented with design.md/ARCHITECTURE.md, ships with tests but no CI. One-day-old burst project.
SparkleButt747 /
Slit-Light-Interference-Sim
School physics project: interactive single-slit light interference simulator using Three.js with custom GLSL shaders, Tweakpane controls, and preset system. Well-documented with structured src/shaders layout but minimal commit history.
SparkleButt747 /
Proce-Tree
Real-time procedural 3D tree generator using L-systems and Three.js. Unfinished experimental project with 15 KB codebase, minimal commit history (1 of 30), created and pushed same day. Has typed config and structured layout but extremely early-stage.
SparkleButt747 /
Galaxy-WebGL
A one-off Three.js galaxy visualization with 60KB codebase, single commit (1 of last 30), untyped JavaScript, no tests/CI. Works but thin scope and minimal trajectory.
SparkleButt747 /
Poem-Master
One-off school project: a simple web app scraping URLs and calling a Heroku backend for poem generation. Vanilla JS/CSS, 3KB, created and pushed same day (2026-03-21). No tests, CI, or meaningful structure.
06 · Timeline
- Feb 27, 2018Joined GitHub
- Mar 15, 2026Created Claudius-Maximus — Claudius Maximus was an experiment at letting Claude iteratively improve itself against the Terminal Bench 2.0 Benchmark; This repo consolidates its findings and allows others to r
- Mar 21, 2026Created RAG
- Mar 21, 2026Created Automaton-QTS — A multi-signal crypto/equity strategy engine combining technical analysis, sentiment fusion, and LLM-assisted decision support with human-in-the-loop oversight.
- Mar 21, 2026Created Galaxy-WebGL — Made a galaxy using ThreeJS, and played around with different random distribution functions to uniformly distribute stars around the galaxy...
- Mar 21, 2026Created Slit-Light-Interference-Sim — This is a simulation of the single slit interference of light, using ThreeJS. Made for a school project; using custom shaders, etc.
- Mar 21, 2026Created Poem-Master
- Mar 21, 2026Created Proce-Tree
- Apr 1, 2026Most recent push to Automaton-QTS
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.