01 · Roasts
The One-Shot Dumper
Advanced-RAG and self-consistency-engine were both born and buried in under a minute — created and last-pushed within seconds of each other. That's not shipping, that's ctrl+V into a repo.
TypeScript Monogamist
97% TypeScript. Every single project. You've discovered one language and you're running it into the ground — admirable loyalty, but the breadth score doesn't lie.
162 PRs, 0 Collaborators
You opened 162 PRs this year and soloPct is 100%. You're holding code reviews with yourself. Your own biggest fan AND harshest reviewer.
Architecture > Tests, Always
harbor has ARCHITECTURE.md, design.md, STATUS.md, AND docs/ — but only 2 repos out of 8 have any tests at all. The docs directory is load-bearing for your quality score.
Speed Runner, No Save Points
persona: 5 days. chaigpt-tools-branching: 1 day. harbor: 2 days. You build entire AI platforms in the time most people take to set up their dev environment, then immediately move on.
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% weight48D
- Consistency20% weight65C
- Quality20% weight72B
- Depth15% weight65C
- Breadth10% weight30F
- Community10% weight40D
03 · Stats
365-day commit heatmap
245 active days
Language distribution
- TypeScript97%
- JavaScript2%
- CSS1%
- HTML0%
- Shell0%
- Dockerfile0%
04 · Numbers
Owned repos
non-fork
14
Commits
last 12 months
1,371
Followers
4
Joined GitHub
Jan 2025
05 · Top repos
harshlocham /
semantask
TypeScript AI task execution platform with monorepo architecture, comprehensive tests, CI/CD, and architectural rigor. Early-stage indie project (2 stars, created Jun 2025) with product iteration evident but limited adoption.
harshlocham /
knowledge-workbench
TypeScript NotebookLM-like RAG app (0 stars); full-stack TanStack Start + React + PostgreSQL + Qdrant + OpenAI; shipped with architecture docs, typed codebase, structured layout, and 30 recent commits across ~2 weeks.
harshlocham /
persona
Educational RAG chatbot grounding Gemini responses in creator content via Qdrant + Next.js. Well-documented system with hexagonal architecture, typed TypeScript, and sophisticated prompt engineering—but pre-launch personal project with 0 stars, no tests, no CI, and only 5 days old.
harshlocham /
chaigpt-tools-branching
TypeScript Next.js chat app with AI tool calling and conversation branching. Two days old, 23 commits, no tests or CI, but typed and architecturally sound with Prisma schema and server actions.
harshlocham /
harbor
TypeScript AI agent SDK with runtime-first loop, provider abstraction, and tool execution. Typed + documented + tested, but brand-new (2 days old) with zero adoption signals.
harshlocham /
harshlocham
GitHub profile config repo with personal README showcasing owner's full-stack projects (Semantask, Knowledge Workbench) and interests. 17 KB, minimal files, 8 commits over 18 months.
harshlocham /
Advanced-RAG
One-shot TypeScript RAG pipeline for ingesting & querying Udemy subtitles with OpenAI embeddings & Qdrant retrieval. No tests, CI, or production signals; brand-new repo (created/pushed same minute).
harshlocham /
self-consistency-engine
TypeScript CLI tool for comparing LLM outputs across OpenAI, Claude, and Gemini. Single commit (1 of last 30), minimal codebase (21 KB), no tests or CI. Early-stage experiment with clear concept and documented setup.
06 · Timeline
- Jan 28, 2025Joined GitHub
- Mar 8, 2025Created harshlocham — Config files for my GitHub profile.
- Jun 28, 2025Created semantask — Originally started as a scalable real-time chat architecture project and evolved into a reliable AI task execution platform.
- Jul 2, 2026Created persona
- Jul 19, 2026Created self-consistency-engine
- Jul 19, 2026Created chaigpt-tools-branching
- Jul 22, 2026Created Advanced-RAG
- Jul 24, 2026Created knowledge-workbench
- Aug 2, 2026Created harbor
- Aug 27, 2026Most recent push to harshlocham
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.