01 · Roasts
The C# Phantom
Your language stats scream 'C# developer' at 89%, but gci-layerjot — the 411 MB elephant in the room — has no README, no license, and no CI. You built a city and forgot to put up street signs.
Sprint Merchant
unvibe: 3 days old. sh1eld: 2 days. blockpins: 3 days. optX: 2 weeks. You have the build velocity of a caffeinated squirrel but the maintenance history of a pop-up shop. Depth requires more than a weekend.
License? Never Heard of Her
rl-fsdp-distillation has ARCHITECTURE.md, STATUS.md, design.md, and a spec — beautiful paperwork — but no LICENSE file. You documented everything except whether anyone is legally allowed to use it.
1 Star Universe
38 repos, 763 commits in a year, 9+ named projects, and the entire portfolio has accumulated exactly 1 star. Use-Anything is holding the entire account's social proof on its back.
CI Allergy
Of 12 scored repos, exactly 2 have CI (kgarg2468 profile repo and aegis). You clearly know how to write tests — 30+ pytest functions in Use-Anything, vitest suites in blockpins — but you refuse to automate them.
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% weight68C
- Consistency20% weight65C
- Quality20% weight59D
- Depth15% weight58D
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
114 active days
Language distribution
- C#89%
- Python4%
- TypeScript3%
- C++2%
- Jupyter Notebook1%
- ShaderLab0%
- Other1%
04 · Numbers
Owned repos
non-fork
30
Commits
last 12 months
763
Followers
17
Joined GitHub
Sep 2024
05 · Top repos
kgarg2468 /
rl-fsdp-distillation
Personal experimental project orchestrating RL + Teacher FT + Distillation pipeline with budget telemetry, schema validation, and audit artifacts. Small codebase (~332 KB) showing structured design but early-stage with 0 stars, no external adoption signals, and no license.
kgarg2468 /
optX
TypeScript/Python full-stack AI business simulator with 6-agent debate system, Monte Carlo + Bayesian networks, and n8n-style node canvas. Typed, documented, structured (75 MB codebase), no tests/CI; ~2 weeks old with 30 commits; experimental stage.
kgarg2468 /
blockpins
Full-stack Next.js pinboard app for Chapman University with Mapbox, Supabase auth, TypeScript, and tested client domain logic. Typed, well-structured, documented, and deployable, but fresh codebase with no external adoption.
kgarg2468 /
Use-Anything
Use-Anything automates agent skill generation from software interfaces via a 5-phase probe-rank-analyze-generate-validate pipeline. Typed Python codebase with structured architecture, comprehensive test suite, and rich docs (README, ARCHITECTURE.md, spec.md), but 1 star/fork indicates early-stage novelty project with n
kgarg2468 /
aegis
AEGIS is a TypeScript/Python cyber defense RL environment with gymnasium gym integration, PPO training pipeline, and evaluation framework. Typed, structured with tests and CI, but lacks README and external adoption signals.
kgarg2468 /
StudySpot
University-specific study spot discovery app with Next.js, Supabase, and Mapbox. Typed, documented, and structured codebase (~158 KB) with multi-step add form, ratings system, and admin dashboard. Early-stage project with no external adoption signals or named products under account.
kgarg2468 /
unvibe
Learning-focused skill bundle for AI coding agents with structured pre-code phase. Has README with clear motivation, installation, and architecture. Typed language unknown, no tests/CI, only 109 KB, 3 days old with 29/30 commits. Experimental foundation stage.
kgarg2468 /
sh1eld
Early-stage Python project for PantherHacks cyber defense demo with tests, structured uv setup, and clear quick-start instructions. No CI, untyped language, single-day development window.
kgarg2468 /
gci-layerjot
Large C# project (411 MB) with test files but no README, docs, CI, or license. Sparse recent commits (3 of last 30 days) suggest sporadic activity. Typed language and meaningful size indicate effort, but lack of documentation and testing infrastructure limit quality assessment.
kgarg2468 /
chapman-data-analytics-club-datathon
Educational datathon project: Streamlit dashboard for energy drink order analysis with KMeans clustering, Plotly visualizations, and OpenAI chat integration. Well-structured for a one-week sprint but limited scope and no external adoption.
kgarg2468 /
ReZone-live
NYC office-to-housing conversion analyzer with Next.js frontend + FastAPI backend scoring feasibility across zoning, utilities, transit, and structural factors. Newly launched (created 2026-03-13), minimally tested, lacks CI/tests but ships typed code and architecture.
kgarg2468 /
kgarg2468
Personal GitHub profile repo with contribution graph generator script. Includes README with hackathon/project links, Python web-scraping script with tests and CI, but untyped code, no license, and minimal standalone functionality.
06 · Timeline
- Sep 10, 2024Joined GitHub
- Jul 10, 2025Created kgarg2468
- Feb 22, 2026Created optX
- Feb 26, 2026Created gci-layerjot
- Mar 13, 2026Created ReZone-live
- Mar 17, 2026Created Use-Anything
- Mar 18, 2026Created chapman-data-analytics-club-datathon
- Mar 21, 2026Created StudySpot
- Mar 31, 2026Created rl-fsdp-distillation
- Apr 4, 2026Created sh1eld
- Apr 5, 2026Created aegis — AEGIS
- Apr 11, 2026Created blockpins
- Apr 15, 2026Created unvibe
- Apr 18, 2026Most recent push to kgarg2468
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.