01 · Roasts
Empty Repo Collector
voiceANALYSIS, GSOC2026, and winshaurya.github.io are three of your repos — combined they have 0 commits, 0 files, and 0 stars. You're farming namespace, not shipping code.
Sprint-and-Ghost
CastroGuard.AI: entire codebase built in 7 hours, agent integrations commented out, never touched again. multiagent-research--system-: 9-day burst, abandoned. You have a great first day, then disappear.
The 850-Line Monolith Manifesto
App.jsx in CastroGuard.AI is 850+ lines of untyped React with no module boundaries. Your portfolio's ty.jsx clocks in at ~2000 LOC. You've discovered components but not the part where you split them up.
TypeScript in Name Only
Your portfolio files have a .tsx extension but HAS_TYPED=no. You renamed your JSX files and called it a day. That's not TypeScript — that's hope with a file extension.
Joined Feb 2025, First Commit Apr 2025
Your heatmap has 4 completely empty weeks at the top — you joined GitHub and then did nothing for two months. At least your most recent push is 2026-05-02, proving you eventually found the keyboard.
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% weight40D
- Consistency20% weight60C
- Quality20% weight62C
- Depth15% weight55D
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
154 active days
Language distribution
- Dart32%
- Jupyter Notebook30%
- JavaScript21%
- TypeScript7%
- HTML5%
- Python3%
- Other2%
04 · Numbers
Owned repos
non-fork
31
Commits
last 12 months
436
Followers
7
Joined GitHub
Feb 2025
05 · Top repos
winshaurya /
toolToParseSaveGFFinBrowser
Proof-of-concept full-stack browser app (FastAPI backend + React + TypeScript + SQLite WASM) for querying genome GFF files in-browser via HTTP range requests and SQL. Typed, documented (README + design.md + ARCHITECTURE.md), multi-file architecture with tests, but no CI and early-stage adoption (0 stars).
winshaurya /
DRI_Survey
Flutter offline-first survey app (family + village) with SQLite local persistence, Supabase cloud sync, and React dashboard. 3.7 MB codebase spans 48 DB migrations, complex multi-table sync logic, and XLSX export—built for field data collection in volatile networks.
winshaurya /
multiagent-research--system-
Early-stage CrewAI-based research assistant generating structured reports from academic sources. Typed Python, documented architecture, modular agent design, but minimal test/CI coverage and ~8 weeks active development (20 commits) indicate experimental status.
winshaurya /
portfolio
A personal portfolio website in React+Vite with polished UI/animations but no documentation, no tests, untyped JavaScript, minimal commit history (4 of last 30), and no real project scope beyond a portfolio showcase.
winshaurya /
CaustMVP
Early-stage research prototype (Jupyter Notebook) demonstrating causal gene intervention for spatial transcriptomics, built for OSRE 2026 proposal. Minimal scope: 6 commits over 2 days, no tests/CI/license, but clear goal and working proof-of-concept with visualizations.
winshaurya /
CastroGuard.AI
Early-stage P&C insurance triage AI dashboard with React UI, LangGraph agent integration, and Leaflet maps. Very new (created 2026-04-14, 7 days old), untyped JavaScript, zero stars/forks, missing README and tests. Demonstrates domain knowledge in underwriting but remains a proof-of-concept.
winshaurya /
jet.Engine.Model
OpenSCAD 3D CAD model repo with zero engagement (0 stars/forks), no documentation, no tests/CI, created and pushed same day in March 2026. Minimal structured output.
winshaurya /
winshaurya.github.io
Empty GitHub Pages scaffold with minimal README and no source files. Created and pushed within seconds (2026-04-25), 1 KB total size, 2 commits. Clear one-shot setup.
winshaurya /
GSOC2026
Empty scaffold repository with zero files, no documentation, no commits since creation. Created on 2026-03-29 with no substantive work or content.
winshaurya /
voiceANALYSIS
Empty scaffold created 2026-03-31 with zero commits, no files, no documentation, and no code substance. Appears to be an uninitialized repository placeholder.
06 · Timeline
- Feb 9, 2025Joined GitHub
- Jan 23, 2026Created DRI_Survey — DRI_Survey APP is FLutter offline-first data ingestion ecosystem engineered for volatile network environments. Its architecture decouples the primary capture layer (Flutter/SQLite)
- Mar 14, 2026Created multiagent-research--system- — A agentic AI research assistant built with CrewAI. Enter any research topic and receive a structured report with accurate citations, extracted evidence, and verified references sim
- Mar 14, 2026Created portfolio
- Mar 22, 2026Created jet.Engine.Model
- Mar 29, 2026Created GSOC2026
- Mar 30, 2026Created toolToParseSaveGFFinBrowser — this browser-extension for a better understing of feature extraction from .gff files using gffutils , that also integrates indexedDB
- Mar 31, 2026Created CaustMVP
- Mar 31, 2026Created voiceANALYSIS
- Apr 14, 2026Created CastroGuard.AI — catastroguard ai is an agentic ai workflow that directly replaces manual p&c insurance triage. it deploys a langgraph pipeline of 10+ ai agents to fetch historical data,
- Apr 25, 2026Created winshaurya.github.io
- May 2, 2026Most recent push to portfolio
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.