01 · Roasts
12-Minute Shipping Strategy
Actual_WPM went from git init to 'shipped' in 12 minutes. That's not an MVP, that's a commit with delusions of grandeur. The Needleman–Wunsch alignment is clever — shame there are zero tests to prove it works.
The Graveyard Architect
53% of your repos are abandoned (staleRepoRatio: 0.53). You ship 7 projects in April, ghost them all in May, then open a new one about mechanistic interpretability. Your repos have a shorter life expectancy than your attention span.
CI? Never Heard of Her
Out of 8 scored repos, exactly 0 have CI. You've got design.md, ARCHITECTURE.md, STATUS.md, and PLAN.md in half your projects — you document like a senior engineer and test like someone's intern.
Hardcoded to Localhost
cs_tripos_partia has '/Users/aryaask/Desktop/' hardcoded throughout. 5 stars, 1 fork, and the README presumably says 'works on my machine' because it literally only works on your machine.
195 Commits, All in Bursts
Your heatmap is 23 weeks of zeros followed by a fireworks display. You don't have a development practice — you have a development emergency. 195 commits/year is fine, but the distribution looks like a seismograph.
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% weight62C
- Consistency20% weight60C
- Quality20% weight67C
- Depth15% weight60C
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
136 active days
Language distribution
- TypeScript77%
- Python9%
- JavaScript7%
- TeX2%
- HTML2%
- CSS1%
- Other2%
04 · Numbers
Owned repos
non-fork
51
Commits
last 12 months
195
Followers
28
Joined GitHub
May 2020
05 · Top repos
AryaaSk /
ratemygithub
TypeScript full-stack GitHub profile scorer with three-pass AI pipeline (file selection, per-repo grading, aggregation), structured codebase, comprehensive rubric system, and production database layer. Early-stage product with 2 stars.
AryaaSk /
residual_stream_visual_decoder
Active research project: visual decoding of LLM residual streams via stroke generation on Gemma 4 E2B. 107MB codebase with comprehensive architecture/training/evaluation docs, multi-stage training pipeline with honest iteration logs, but no CI/tests/license and 3-day age limit exploration scope.
AryaaSk /
quant
Multi-market transformer trading POC with LLM-scraped data and walk-forward backtests. Comprehensive infrastructure with honest limitations documentation; core model collapsed to constant predictor. Well-documented experimental system with property-based temporal leak testing and realistic execution costs.
AryaaSk /
Zootropolis
Early-stage TypeScript AI automation project with comprehensive architectural documentation (design.md, ARCHITECTURE.md, STATUS.md), test files, but no CI pipeline or traditional README. Created April 15, 2026 with 30 commits in 2 days—rapid burst prototype.
AryaaSk /
replay
Ambitious pre-release macOS app (Tauri + Rust + TypeScript sidecar) turning screen recordings into AI-ingestable bug reports via Claude vision. Typed, documented (README + ARCHITECTURE.md + BUILDING.md + PLAN.md), structured multi-layer codebase (~7.7 MB), but barely shipped (1 star, created 2 days ago, 30 commits), no
AryaaSk /
cs_tripos_partia
Personal Cambridge exam prep system with AI question generation (Python + Claude API). Well-documented system design, typed Python code, structured knowledge map, but minimal public adoption (5 stars, no external use).
AryaaSk /
marginal_edge
Educational framework for building betting prediction pipelines with a worked Liverpool example. TypeScript + Next.js, comprehensive type system, test suite, but brand-new (3 days old, 4 commits) and minimal external adoption (3 stars).
AryaaSk /
Actual_WPM
Freshly-launched typing test app with LLM-assisted accuracy scoring. TypeScript + Next.js 16, coherent novel concept, but minimal deployment history (4 commits in <1 hour), no tests/CI, and zero adoption signals yet.
06 · Timeline
- May 3, 2020Joined GitHub
- Apr 2, 2026Created cs_tripos_partia — Agent to help prepare for CS Tripos Part IA exams
- Apr 11, 2026Created Actual_WPM — stop fixing your typos. the llms already know what you meant
- Apr 15, 2026Created Zootropolis — Full AI run companies
- Apr 18, 2026Created ratemygithub — Rate your GitHub against others!
- Apr 27, 2026Created replay — Click record. Show your bug. Get a perfect description for your AI agent.
- May 14, 2026Created marginal_edge — Educational testbed: can your model beat a betting market? Build a pipeline, train a predictor, simulate the bets.
- May 16, 2026Created quant — Multi-market transformer trading POC with agent-scraped text and Voyage embeddings
- May 19, 2026Created residual_stream_visual_decoder — Visual lens into LLM residual streams: stroke-output decoder of Gemma 4 E2B activations, NLA-style autoencoder with vision-pathway reconstruction. Research project.
- May 22, 2026Most recent push to residual_stream_visual_decoder
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.