01 · Roasts
Burst-and-Ghost Engineer
8 repos, all built in 1–13 days, then abandoned. knownissue is 7 weeks old and already your longest-maintained project. You don't write software, you sketch software.
Credential Leak Artist
sentinel-web hardcodes Supabase keys and pricing config directly in source. You're building an AI *governance* product while committing secrets to GitHub. The irony is load-bearing.
The 10-Minute Man
foodfinder's own repo description says '10 minute project'—and the commit timestamps confirm it: created 20:04, last push 20:09. At least you're honest about the effort level.
Solo Forever
soloPct = 100%. 22 public repos, 8 followers, 0 forks, 0 contributors. You've built an entire portfolio in complete isolation. GitHub is your diary, not your front door.
Heatmap Desert with an Oasis
Weeks 1–32 of your heatmap are nearly all zeros. Then weeks 33–47 explode with 4s. You don't have a consistency habit—you have a deadline panic reflex.
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% weight65C
- Quality20% weight72B
- Depth15% weight58D
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
99 active days
Language distribution
- TypeScript57%
- C++33%
- Python2%
- HTML2%
- CSS1%
- JavaScript1%
- Other4%
04 · Numbers
Owned repos
non-fork
21
Commits
last 12 months
611
Followers
8
Joined GitHub
Feb 2022
05 · Top repos
gong8 /
knownissue
TypeScript monorepo (Turborepo + pnpm) providing MCP server for agent-driven bug/patch tracking with credit economy. Ships typed, tested, CI'd—well-structured indie project with named product and clear use case.
gong8 /
3Roads
TypeScript quiz bowl question generator monorepo with API server, React frontend, MCP tools, and Prisma database. Young project (5 days old) with structured multi-package layout, typed code, and meaningful documentation via README + ARCHITECTURE.md, but no tests, CI, or license yet.
gong8 /
willow-demo
Knowledge graph system for AI memory combining Rust NAPI backend with React/Hono frontend. Ships typed, documented, tested project with complex architecture; fresh repo (4 days old) but substantial engineered artifact with clear vision.
gong8 /
nasty-plot
Young Pokemon team-building & analysis tool (Turbo monorepo, Next.js 16, Prisma, LLM-powered). Typed, tested, multi-package architecture with domain logic (type charts, damage calc, battle AI). Limited adoption (1 star, 8 days old) but non-trivial shipped project.
gong8 /
cramkit
Local-first exam study tool (MCP server + React frontend + SQLite backend) integrating Claude with PDF processing, knowledge graph extraction, and chat. Typed, documented, tested, structured monorepo built in ~3 days; clear experimental scope but working implementation.
gong8 /
sentinel-web
TypeScript landing page for Sentinel AI governance product with polished React components, Stripe/Resend integration, and design partner form—zero stars, no documentation, no tests or CI, clear early-stage SaaS marketing site.
gong8 /
nameguesser
Minimalist Wordle-inspired name guessing game in TypeScript + React with Tailwind styling. Created today (2026-01-29) with 1 commit; typed, documented via README, structured src/ layout with game logic and components; no tests or CI. Clearly a completed personal project shipped in a single burst.
gong8 /
foodfinder
10-minute TypeScript web scraper + Claude integration for food recommendations. Typed, structured layout, but no README, tests, CI, or documentation. Single-burst commit within hours of creation.
06 · Timeline
- Feb 1, 2022Joined GitHub
- Jan 17, 2026Created sentinel-web — sentinel landing page
- Jan 29, 2026Created nameguesser — guess properties of people's names!
- Feb 9, 2026Created nasty-plot — Pokemon team building tool and competitive analysis playground. Build teams, analyze matchups, calculate damage, and learn competitive Pokemon concepts.
- Feb 16, 2026Created cramkit — simple local mcp server to enhance context for exam cramming
- Feb 19, 2026Created willow-demo — knowledge graphs for AI memory, demo version
- Mar 2, 2026Created foodfinder — vibe-coded, claude-cli based web scraper to recommend food based off location and mood (10 minute project)
- Mar 8, 2026Created knownissue — Shared debugging memory for AI coding agents. Search, report, patch, and verify issues via MCP. Every fix survives beyond the conversation.
- Mar 20, 2026Created 3Roads — trivial quizbowl-style question generated, uses claude-cli
- Apr 24, 2026Most recent push to knownissue
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.