01 · Roasts
Test suite: missing in action
All 12 scored repositories report HAS_TESTS=no. The shipping velocity is real; the regression safety net is fictional.
CI is a rare spawn
Only devwez and spark have CI. DeepDNS, sortforge, and waftsolutions carry serious functionality without automated validation.
Horizontal builder energy
92 recent commit samples and 37 public repositories say prolific; several single-file projects say the architecture arrived after the deadline.
Adoption has a flagship
DeepDNS has 41 stars and graphyte has 34; the rest of the portfolio is still waiting for its crowd to show up.
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% weight80A
- Quality20% weight57D
- Depth15% weight55D
- Breadth10% weight80A
- Community10% weight25F
03 · Stats
365-day commit heatmap
255 active days
Language distribution
- Python75%
- TypeScript15%
- Go3%
- HTML3%
- CSS2%
- JavaScript1%
- Other1%
04 · Numbers
Owned repos
non-fork
16
Commits
last 12 months
1,799
Followers
5
Joined GitHub
Jul 2022
05 · Top repos
devwez /
DeepDNS
DeepDNS is a documented, MIT-licensed Go/React DNS firewall with Docker deployment, SQLite analytics, WebSocket live events, and a 200K+ domain blocklist, but lacks tests and CI.
devwez /
waftsolutions
A polished, typed React/Vite studio landing page with six service cards, four pricing tiers, animated sections, and resilient geo-currency detection, but limited adoption and no tests or CI.
devwez /
orrery
A polished, interactive browser orrery with real NASA-derived orbital data, Newton–Raphson visualization, editing controls, and a deployed Vercel demo, but limited adoption and no tests, CI, license, or typed code.
devwez /
sortforge
A documented TypeScript/React and Go sorting benchmark with 15 algorithms, Web Worker execution, Canvas visualization, Monaco playground, and SQLite leaderboard; its small audience and apparent frontend compile defects limit current maturity.
devwez /
spark
A polished, documented Vite/React/TypeScript starter with strict configuration and CI, but only 2 stars, no tests, and a small template-sized codebase limit demonstrated adoption and depth.
devwez /
devwez
A small GitHub profile repository that automates a generated metrics.svg via scheduled and push-triggered GitHub Actions; it has minimal documentation and no application source or tests.
devwez /
graphyte
A documented, MIT-licensed Python utility that maps a 7x52 pattern to dated Git commits, but it remains a small single-script project without tests, CI, typing, or broad adoption.
devwez /
keydungeon
A playable, documented keyboard-only roguelike in a single index.html, with procedural rooms, combat, enemy AI, fog of war, five floors, and local best-run persistence; adoption remains minimal at 2 stars.
devwez /
karoo-weave
A polished single-page HTML landing site for a South African sustainable-fashion concept, with responsive sections, accessibility touches, deployment metadata, product messaging, and a cited SOURCES.md, but no demonstrated adoption or sustained repository history.
devwez /
devdash
A compact single-file HTML dashboard with a polished project catalog and localStorage-backed todo list, but no documentation, tests, CI, typed code, or evidence of sustained repository evolution.
devwez /
soundpost
A documented single-file browser prototype for transmitting text as audio, but the implementation is incomplete and has no adoption, tests, CI, or license artifact.
devwez /
tagless
Empty repository scaffold with no source files, documentation, tests, CI, license, or stated project purpose.
06 · Timeline
- Jul 6, 2022Joined GitHub
- Feb 13, 2026Created graphyte — Render custom patterns and text onto your GitHub contribution graph using automated commits.
- Jun 27, 2026Created devwez — Weza Mwiwa — software builder from South Africa. SortForge, Graphyte, DeepDNS, Hack Club projects.
- Jun 30, 2026Created DeepDNS — Network-wide ad & tracker blocking DNS server. Go backend, React dashboard, 200K+ blocked domains, real-time analytics via WebSocket.
- Jul 5, 2026Created sortforge — Visual sorting algorithm benchmark. 15 algorithms, Web Worker parallel execution, canvas visualization, Monaco editor playground.
- Jul 22, 2026Created spark — A production-ready Vite + React + TypeScript starter with Tailwind CSS, dark mode, and developer tooling.
- Jul 29, 2026Created waftsolutions — Waft Solutions — premium web development studio. Global pricing that adapts to your region.
- Aug 4, 2026Created orrery — Live solar system in the browser — real Keplerian orbits (Newton-Raphson), no frameworks, no build step
- Aug 21, 2026Created keydungeon — Keyboard-only roguelike dungeon crawler — no mouse, no tab, arrow keys only
- Aug 27, 2026Created karoo-weave — Karoo Weave — From Earth to Wardrobe. South Africa's sustainable fashion brand.
- Aug 29, 2026Created devdash
- Sep 9, 2026Created soundpost — sends text through the air as sound. no internet, no bluetooth, just a speaker and a mic
- Sep 25, 2026Created tagless
- Sep 25, 2026Most recent push to tagless
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.