01 · Roasts
Zero Stars, Maximum Ambition
21 repos, 0 total stars across all of them. You've shipped a PyPI package (TraceRazor), a bilingual face-recognition kindergarten tool (Eton_Vision), and a multi-agent fact-checker — and somehow not a single outsider has clicked ⭐. Marketing budget: $0.
CI? Never Heard of Her
8 out of 9 scored repos have HAS_CI=no. TraceRazor is the lone adult in the room with its cargo check + clippy workflow. The rest are living on prayer and local `npm run dev`.
HTML 53%: Not What You Think
Your language breakdown says you're half an HTML developer. Probably static assets and notebooks inflating the count — but from the outside it just looks like you're very enthusiastic about angle brackets.
The 36-Hour Architect
AgenticAgentAgenting: created 2026-05-16, last push 2026-05-17. FastAPI + LangGraph + Next.js + Playwright tests + ARCHITECTURE.md, all in roughly 36 hours. Either extremely productive or you asked an LLM to cosplay as a senior engineer.
4 Followers, 19 PRs
You opened 19 pull requests this year on other people's repos but have only 4 followers. You're out here contributing to the community like a ghost — no one knows you exist yet.
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% weight67C
- Depth15% weight65C
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
109 active days
Language distribution
- HTML53%
- Jupyter Notebook25%
- TypeScript10%
- Python4%
- JavaScript4%
- Rust2%
- Other2%
04 · Numbers
Owned repos
non-fork
19
Commits
last 12 months
322
Followers
4
Joined GitHub
Jul 2020
05 · Top repos
ZulfaqarHafez /
TraceRazor
TraceRazor: framework-agnostic Rust+Python analyzer for AI agent token efficiency. Computes 13 structural metrics (SRR, LDI, TCA, etc.) in <5ms offline; ships with tests, CI/CD, typed Rust core, comprehensive docs (design.md, ARCHITECTURE.md), and working CLI auditing tool. Early-stage but coherent product solving real
ZulfaqarHafez /
OsciWriting
Rigorous LLM-caching measurement study with comprehensive design doc, typed Python code, structured multi-file architecture (10+ modules), full test suite, and cost-model derivation. Pre-decision research project, not yet published/deployed—intended as internal evidence to guide downstream project commitment.
ZulfaqarHafez /
Eton_Vision
TypeScript React app for early childhood observations with face recognition, AI report generation, and bilingual support. Well-structured, documented, and typed but pre-launch (0 stars, personal project).
ZulfaqarHafez /
AAI3008_ClaimLens
Agentic fact-checking system with LangGraph orchestration, fine-tuned NLI verification, and real-time Next.js frontend. Academic project (AAI3008) with multi-agent decomposition pipeline, but no CI/license, limited tests, and untyped Python.
ZulfaqarHafez /
Our_World
Private couple's web app with React frontend, Express/Supabase backend, and AI study assistant. Typed, documented, ~2.2 MB codebase with structured architecture but no public adoption or team output beyond this single project.
ZulfaqarHafez /
AgenticAgentAgenting
Experimental multi-agent agentic system with FastAPI + LangGraph + Next.js. Shipped with tests, CI-less, untyped Python backend. New repo (2 days old, 4 recent commits). Demonstrates architectural breadth but early-stage quality gaps.
ZulfaqarHafez /
portfolio_website
Personal portfolio website built with React 19, TypeScript, and Tailwind CSS. Fully featured with dark mode, animations, analytics integration, and structured component architecture. Well-typed and documented, but lacks tests and CI/CD—typical of a shipped portfolio project without external adoption.
ZulfaqarHafez /
Marquee
TypeScript Next.js web app for AI-powered webtoon panel restyling with multi-model orchestration, audited self-correction, and animation export. Typed, structured, documented but brand-new with zero adoption signals.
ZulfaqarHafez /
Eton-House
TypeScript/React educational tool with AI-powered report generation; minimal README, no tests visible in samples, no CI/license, but typed+documented structure indicates active development over ~7 weeks.
06 · Timeline
- Jul 13, 2020Joined GitHub
- Jan 5, 2026Created portfolio_website — Portfolio_website for myself
- Jan 14, 2026Created Eton-House — Bilingual Group Progress Report
- Feb 4, 2026Created AAI3008_ClaimLens — Agentic fact-checking pipeline with fine-tuned DeBERTa-v3 NLI, powered by LangGraph. Decomposes text into atomic claims, retrieves web evidence, and verifies with confidence scorin
- Feb 24, 2026Created Our_World — Our world
- Mar 2, 2026Created Eton_Vision
- Apr 4, 2026Created TraceRazor — TraceRazor is a framework-agnostic, high-performance auditor for AI agent reasoning traces. It analyses completed or in-flight agentic execution paths and produces actionable effic
- May 9, 2026Created Marquee
- May 16, 2026Created AgenticAgentAgenting
- May 19, 2026Created OsciWriting
- May 23, 2026Most recent push to Eton_Vision
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.