01 · Roasts
Burst-and-Ghost Developer
WikiDream: 30 commits in 1 day. Nexus: 2 commits in 5 minutes. NanoMathWeb: entire history in under 1 hour. You've perfected the art of creating a GitHub repo, panic-coding everything in a single sitting, and never opening it again.
CI? Never Heard of Her
9 repos, exactly 1 has CI — and it's the Python scraper (Jobert). Your TypeScript monorepo (DevClaw), your LLM (NanoMath), your Swift app (Nexus): all flying blind with zero automated quality gates. You clearly know what CI is; you just refuse to use it.
Professional Coursework Shipper
MidasCore is explicitly 'completed as part of the J.P. Morgan Chase Forage experience' and Streem is 'A-Level coursework.' Nothing wrong with learning, but calling yourself a systems developer while 2 of your top repos are homework assignments is a bold choice.
The 8-Month Ghost
Your heatmap shows 35 consecutive weeks of zero activity before suddenly going full-throttle in week 37. Whatever happened between January and September 2025, the GitHub servers were not involved.
5 Total Stars Across 37 Repos
DevClaw converts tasks to GitHub PRs via WhatsApp, NanoMath trains a 136M parameter LLM, and somehow the entire portfolio has accumulated 5 stars. You're building in a sealed room and forgetting to open a window.
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% weight60C
- Quality20% weight67C
- Depth15% weight55D
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
118 active days
Language distribution
- TypeScript69%
- Dart6%
- Python5%
- Swift5%
- HTML3%
- JavaScript3%
- Other9%
04 · Numbers
Owned repos
non-fork
36
Commits
last 12 months
221
Followers
10
Joined GitHub
Jan 2021
05 · Top repos
agmada-asa /
DevClaw
Multi-agent AI system converting plain-English tasks to GitHub PRs via WhatsApp/Telegram, built with TypeScript, modular microservices (openclaw-gateway, orchestrator, agent-runner), comprehensive architecture docs, and Turborepo monorepo structure. Demonstrates non-trivial scope with structured multi-service coordinat
agmada-asa /
NanoMath
Early-stage GPT-style LLM project with end-to-end training pipeline. Ships typed Python code with custom tokenizer, synthetic data generation (1M math problems), and inference CLI. No tests, CI, or license. Created and pushed within 6 days; demonstrates sustained focused work but repo is nascent.
agmada-asa /
WikiDream
Fresh Wikipedia reader UI leveraging Next.js App Router with typed code, structured layout, and live Vercel demo. Minimal commit history and zero adoption signals limit impact; early-stage but well-intentioned execution.
agmada-asa /
Jobert
Functional serverless job scraper for internships, deployed via GitHub Actions with Telegram notifications. No tests, untyped Python, but well-documented README with clear setup instructions and working CI/CD automation.
agmada-asa /
TAPE
Podcast clip generator desktop app using Whisper + Llama 3.2 with Tkinter GUI. Typed Python with clear architecture, meaningful README, structured layout (main.py, generateIdeas.py), but lacks tests, CI, and production signals (0 stars/forks).
agmada-asa /
Streem
A-Level coursework audio streaming platform with React frontend, Node.js backend, and Python ML pipelines for recommendations and lyric generation. TypeScript across client and server, Firebase integration, clean architecture but minimal production adoption signals and recent burst development.
agmada-asa /
Nexus
A Swift/Node.js local AI chat app with RAG capabilities, created 2/21/26. Clean, typed codebase with solid architecture (SwiftUI frontend, Express backend, LlamaIndex RAG), comprehensive README, but only 2 commits in 5 minutes suggests minimal sustained development.
agmada-asa /
NanoMathWeb
Early-stage Next.js + TypeScript web frontend for NanoMath LLM with Tailwind CSS UI. Typed, documented, and deployed, but no tests/CI and minimal commit history (5 of 30 in sampling window) indicates fresh prototype.
agmada-asa /
MidasCore
J.P. Morgan Forage coursework: Java 17 Spring Boot event-driven transaction system with Kafka, JPA persistence, and integration tests. One-shot educational project with no external adoption or impact.
agmada-asa /
agmada-asa
GitHub profile README with no functional code, no tests/CI, 0 stars, 3 commits. Single-file personal landing page referencing other projects but containing no executable software.
06 · Timeline
- Jan 16, 2021Joined GitHub
- Dec 13, 2024Created TAPE
- Feb 20, 2026Created NanoMath
- Feb 21, 2026Created Nexus
- Feb 21, 2026Created Streem
- Feb 21, 2026Created agmada-asa
- Feb 24, 2026Created WikiDream
- Feb 26, 2026Created NanoMathWeb
- Mar 1, 2026Created DevClaw
- Mar 9, 2026Created MidasCore
- Mar 14, 2026Created Jobert
- Apr 24, 2026Most recent push to Jobert
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.