▸ This tool was built by an AI agent from Zoral
← RATE MY GITHUB

#16 — Top 99.0%

twangodev

James Ding

A

Ship machine

Overall

0.0

/ 100

01 · Roasts

289 PRs, 2 Issues — Are You Okay?

You opened 289 pull requests this year but only 2 issues. Either you never discover bugs until you've already fixed them, or you're merging your own PRs before anyone else notices the repo exists.

The Graveyard Starter Pack

`status` repo: created and last-pushed within the same *second* on Aug 25. Zero files. Zero docs. One commit. At least rename it `aspirations` and call it performance art.

91 Total Stars Across 73 Repos

That's 1.25 stars per repo on average. You're shipping FFmpeg pipelines, HDR mergers, WASM libraries, and JetBrains plugins — and the internet is awarding you a participation trophy. The craftsmanship is real; the audience just hasn't found you yet.

soloPct: 96%

96% of your commits are solo. You built a JetBrains IDE plugin, a Rust image alignment library, and a speech diffusion model — entirely alone. Either you don't need help or you've never asked. Touch grass (and a collaborator).

New Repo Every Other Week

speech-dllm (5 days old), librivox-mirror (3 days old), and postframe all shipped in August alone. You commit 3391 times a year and start projects faster than `cargo init`. Finishing them at that same pace would be a different story.

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

  • Impact
    25% weight
    68C
  • Consistency
    20% weight
    95S
  • Quality
    20% weight
    79B
  • Depth
    15% weight
    68C
  • Breadth
    10% weight
    80A
  • Community
    10% weight
    55D

03 · Stats

365-day commit heatmap

340 active days

Less
More

Language distribution

6 langs
  • Python41%
  • TypeScript21%
  • Rust18%
  • Svelte11%
  • Go7%
  • Kotlin2%

04 · Numbers

Owned repos

non-fork

65

Commits

last 12 months

3,391

Followers

40

Joined GitHub

Mar 2019

05 · Top repos

twangodev /

sdocx

68/100

Reverse-engineered reverse-engineering toolkit for Samsung Notes `.sdocx` format; Rust library + CLI + WASM, typed with strong docs, CI/tests, and architecture evident from design.md + notebooks—niche but non-trivial shipped project showing sustained effort across 6 months (19 watchers, published to crates.io and npm).

I55Q82D68
READMECITyped
Rust19this week

twangodev /

jetplay

67/100

JetBrains IDE plugin enabling audio/video playback via JCEF and FFmpeg, with comprehensive format support, transcoding, and media metadata extraction. Solid indie SaaS-style project with named product distribution.

I55Q80D65
READMETestsCITyped
Kotlin7this week

twangodev /

twangodev

55/100

Personal portfolio site (twango.dev) built with SvelteKit, TypeScript, and Tailwind. Showcases blog, timeline visualization, flight tracking, and GPG key management—demonstrates solid architectural depth with 98 KB of organized TypeScript modules.

I40Q65D60
READMETestsCITyped
Svelte0this week

twangodev /

gmetrics

55/100

A well-structured Go port of the metrics SVG generator with full typed architecture, comprehensive tests, CI/CD, and documentation. Active but young project (~3 months old) with 30 commits, showing solid craftsmanship but early adoption.

I40Q75D50
READMETestsCITyped
Go07d ago

twangodev /

mtb-align

52/100

A highly polished, well-tested Rust library implementing Ward's median-threshold-bitmap algorithm for aligning handheld HDR exposures. Strong craftsmanship with comprehensive tests and CI, but minimal adoption (1 star) limits impact.

I25Q80D50
READMETestsCITyped
Rust112d ago

twangodev /

postframe

50/100

Specialized Rust photography tool with strong technical craft—types, tests, CI—but nascent ecosystem adoption. Emerged Aug 2026, zero stars but well-architected; CLI + web UI + desktop shipping real HDR merging pipeline.

I25Q75D50
READMETestsCITyped
TypeScript0this week

twangodev /

airwer

48/100

Specialized ATC Word Error Rate library with domain-specific normalizations (callsigns, frequencies, flight levels). Well-typed, thoroughly tested, and published to PyPI, but niche use case and zero external adoption so far.

I25Q72D50
READMETestsCI
Python0this week

twangodev /

speech-dllm

45/100

Very recent (5 days old) Python research project on diffusion language models for speech with typed codebase, tests, README, and solid architecture but minimal adoption or external validation.

I25Q60D35
READMETests
Python0this week

twangodev /

librivox-mirror

42/100

Actively developed Python tool for mirroring LibriVox audiobooks to Hugging Face datasets, with structured architecture, comprehensive testing, and CI/CD. Early-stage project (created Aug 27, 2026) with 0 stars but published on PyPI, showing genuine portfolio work rather than experimental code.

I25Q60D35
READMETestsCI
Python0this week

twangodev /

skytrack

40/100

Niche Hypixel Skyblock market tracker with typed Svelte frontend, D1 database backend, comprehensive price tracking logic, and scheduled workers. Zero stars and recent 2026 timestamp suggest active development or future repo.

I15Q60D45
READMETestsCITyped
Svelte0this week

twangodev /

cohere-transcribe-cli

35/100

Personal CLI tool wrapping Cohere Transcribe model with PyPI distribution. Python project with README and MIT license, but no tests, CI, or type hints. Modest scope (~195 KB) with 10 commits in ~4 months.

I25Q45D35
README
Python0this week

twangodev /

status

2/100

Empty scaffold repo with no files, no documentation, and single initial commit. Not a functional project.

I5Q0D5
Unknown07d ago

06 · Timeline

  1. Mar 23, 2019
    Joined GitHub
  2. Jun 15, 2021
    Created skytrack — The leading Hypixel Skyblock price tracker. Stay updated with market trends, item values, and trade insights.
  3. Aug 7, 2023
    Created twangodev — A carefully curated illusion that I have my shit together.
  4. Mar 7, 2026
    Created sdocx — Reverse-engineered tooling and SDK for converting Samsung Notes (.sdocx) files.
  5. Mar 30, 2026
    Created jetplay — Audio and video playback for JetBrains IDEs
  6. Apr 30, 2026
    Created cohere-transcribe-cli — A simple CLI for Cohere Transcribe.
  7. May 22, 2026
    Created gmetrics — A Go port of lowlighter/metrics
  8. May 29, 2026
    Created airwer — Word Error Rate for Air Traffic Control
  9. Jul 30, 2026
    Created mtb-align — Median threshold bitmap alignment for handheld HDR exposures
  10. Jul 30, 2026
    Created postframe — Post-processing built on your JPEGs
  11. Aug 24, 2026
    Created speech-dllm — Diffusion language models for speech
  12. Aug 25, 2026
    Created status — some status pages
  13. Aug 27, 2026
    Created librivox-mirror — A continuously updated LibriVox audio mirror.
  14. Aug 29, 2026
    Most recent push to twangodev

07 · Compare

github.com/
twangodev · 6dmedian coder

08 · Rubric

How this score was produced

Overall = Σ (category × weight) + gentle top-end curve

CategoryWeightScoreContrib.
Raw total75.5
Top-end curve+5.6
Final overall81.1

Tier thresholds

S90100Mass-producing humansA8089Ship machineB7079Solid engineerC6069Getting thereD4059README enthusiastF039GitHub tourist
▸ How the pipeline works
  1. 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.
  2. 02Triage.A small model reads every repo's file tree + README and picks the 20 files per repo that actually reveal how you code.
  3. 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.
  4. 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.
  5. 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.
twangodev · 81.1/100 — Rate My GitHub