01 · Roasts
The 39-Minute Architect
ideation-workflow was born and abandoned within a single lunch break — committed at 21:18, last push at 21:57, 39 minutes of ambition before the tab was closed forever. The README describes a 5-stage workflow system; the repo delivers empty output/ folders.
Star Inequality
43 of your 48 total stars live in one repo (eduba-brand), a design system that's essentially a README with tokens. The other 5 repos share the remaining 5 stars. Your documentation ships harder than your code.
Heatmap Archaeology
The first 11 weeks of your heatmap are a perfect void. Then a burst. Then another void. Then another burst. Your commit history looks less like a developer's schedule and more like a seismograph in a geologically quiet region.
Solo Artist
soloPct = 95%, totalPRsYear = 1, totalIssuesYear = 0. You have 17 followers and follow 2 people. GitHub is clearly a private journal for you — well-formatted, occasionally interesting, but mostly a conversation with yourself.
CI/CD Allergist
Only workloom has CI. Out of 6 scored repos, 5 have no CI pipeline. 4 have no tests. You're building AI agents, design systems, and SaaS products on a foundation of 'it works on my machine' theology.
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% weight58D
- Consistency20% weight55D
- Quality20% weight57D
- Depth15% weight50D
- Breadth10% weight65C
- Community10% weight40D
03 · Stats
365-day commit heatmap
79 active days
Language distribution
- TypeScript42%
- JavaScript20%
- SCSS12%
- Python11%
- Kotlin6%
- HTML4%
- Other5%
04 · Numbers
Owned repos
non-fork
21
Commits
last 12 months
288
Followers
17
Joined GitHub
Feb 2020
05 · Top repos
banbury-cheese /
eduba-brand
Eduba brand design system repo with comprehensive documentation (AGENT_BRIEF.md, BRAND.md, PRODUCT_UI.md, tokens/, animations/, assets/) structured for AI agents. 43 stars, 220 KB, 6 commits over 5 days; ships Agent Skill at .agents/skills/. No tests or CI, untyped JS.
banbury-cheese /
workloom
Local-first desktop activity tracker with Ollama vision and LLM digests. Early-stage personal project (1 star, created April 2026) with typed Python, extensive architecture, and working MVP spanning multiple subsystems.
banbury-cheese /
eduba.io
Early-stage Next.js + Python agent SaaS for AI-driven marketing page generation. TypeScript frontend with Sanity CMS integration; Python backend orchestrates web discovery, multi-source ingestion, and OpenAI-powered sector content creation. Typed, structured, documented via README; no tests or CI. Created Jan 2026, 30
banbury-cheese /
llm-games
TypeScript Next.js 15 learning game app with 13 game modes, AI-powered deck generation via Vercel AI SDK, localStorage persistence, and comprehensive documentation (README, docs/ folder, ARCHITECTURE.md, STATUS.md). Lacks tests, CI, and license; 21 of 30 recent commits suggests active development within 2 weeks.
banbury-cheese /
itskay.co
Personal portfolio site built as a retro macOS desktop UI in Next.js. Typed TypeScript, tested with Playwright and Vitest, structured with clean store (Zustand) and component architecture. Created within one week with minimal commits—a polished one-off project.
banbury-cheese /
ideation-workflow
Single-day ideation workflow template using folder-based AI agent architecture. No code, minimal file tree, zero commits beyond initial push. README documents a 5-stage process but lacks executable depth or adoption signal.
06 · Timeline
- Feb 2, 2020Joined GitHub
- Jan 20, 2026Created eduba.io
- Feb 24, 2026Created llm-games
- Feb 25, 2026Created itskay.co
- Mar 12, 2026Created ideation-workflow
- Apr 13, 2026Created eduba-brand
- Apr 13, 2026Created workloom — Local-first desktop activity tracker with Ollama vision, sessionization, and daily LLM digests
- Apr 27, 2026Most recent push to eduba.io
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.