▸ This tool was built by an AI agent from Zoral
← RATE MY GITHUB

#104 — Top 93.3%

Json604

Kartikey Attri

B

Solid engineer

Overall

0.0

/ 100

01 · Roasts

Portfolio, not audience

You shipped at least six named projects, including live Spend, but totalStars is still 0.

CI is the missing teammate

scrapeverse has Actions, while spend_native, silentbug-bench, and genAI_assignment still ship without CI.

Benchmark assembly line

silentbug-bench, sycophancy-bench, and rewardhack-bench collectively define 78 benchmark tasks; now get humans to run them.

Horizontal builder

165 multi-repo recent commits and 444 yearly commits say you build broadly; sustained public adoption has not caught up.

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

  • Impact
    25% weight
    68C
  • Consistency
    20% weight
    65C
  • Quality
    20% weight
    79B
  • Depth
    15% weight
    58D
  • Breadth
    10% weight
    80A
  • Community
    10% weight
    25F

03 · Stats

365-day commit heatmap

76 active days

Less
More

Language distribution

7 langs
  • TypeScript40%
  • Python23%
  • JavaScript14%
  • Kotlin6%
  • Java5%
  • CSS4%
  • Other8%

04 · Numbers

Owned repos

non-fork

30

Commits

last 12 months

444

Followers

5

Joined GitHub

Sep 2024

05 · Top repos

Json604 /

scrapeverse

62/100

Driftwatch is a substantial typed TypeScript web-monitoring system with a tested break-vs-change classifier, guarded healing saga, multi-source adapters, dashboard, CLI, and scheduled Bright Data ingestion; adoption is not yet evidenced.

I20Q82D50
READMETestsCITyped
TypeScript012d ago

Json604 /

spend_native

59/100

Spend is a substantial typed Android expense tracker with React Native/Kotlin clients, Fastify/Postgres sync, SMS parsing, self-hosted updates, and strong transactional/conflict-handling tests, but it has no stars, CI, or license.

I55Q68D50
READMETestsTyped
TypeScript022d ago

Json604 /

silentbug-bench

58/100

A documented 18-task PyTorch benchmark with gold patches, hidden defect tests, deterministic generation, and an out-of-process verifier; strong engineering scope but no demonstrated adoption or CI.

I20Q75D50
READMETests
Python015d ago

Json604 /

genAI_project

48/100

A deployed, typed Next.js multimodal catalogue search with 400 products, Jina/Gemini enrichment, Supabase pgvector retrieval, analytics, and validation; adoption remains unproven at 0 stars.

I35Q68D35
READMETestsTyped
TypeScript02mo ago

Json604 /

genAI_assignment

44/100

A substantial coursework portfolio with four GenAI assignments, including a deployed persona chatbot, typed RAG modules, and a feature-rich Playwright/CDP web automation agent, but no tests, CI, license, or visible adoption.

I25Q58D50
README
JavaScript02mo ago

Json604 /

rewardhack-bench

38/100

A documented, test-backed Python benchmark with 20 deterministic reward-hacking tasks, a JSON-only XPS interpreter, executable verifier, and six exploit-search strategies; adoption is not yet evidenced by its 0 stars and 0 forks.

I22Q52D35
READMETests
Python015d ago

Json604 /

hld_typeahead

35/100

A well-documented FastAPI typeahead service with trie indexing, consistent-hash caching, trending scores, SQLite batching, and a substantial pytest suite, but it is a same-day, zero-star project without CI or a license.

I25Q50D25
READMETests
Python02mo ago

Json604 /

sycophancy-bench

34/100

A carefully engineered Python benchmark with 40 paired flaw/control tasks, deterministic synthetic experiments, framing probes, and strict scoring; adoption is not yet evidenced by its 0 stars and same-day launch.

I20Q58D25
READMETests
Python015d ago

Json604 /

devops-assignments

32/100

A well-documented DevOps coursework repository spanning seven assignments, with reproducible shell evidence and Docker experiments, but no tests, CI, license, or demonstrated external adoption; README also records a broken Python Dockerfile.

I20Q40D35
README
Shell0this week

Json604 /

Json604

8/100

A 4 KB profile-style repository with a GitHub README and social links, but no fetched source files, tests, CI, license, or demonstrated software output.

I10Q10D5
README
Unknown028d ago

06 · Timeline

  1. Sep 29, 2024
    Joined GitHub
  2. Mar 8, 2026
    Created Json604
  3. Apr 29, 2026
    Created genAI_assignment
  4. Jun 14, 2026
    Created genAI_project
  5. Jun 21, 2026
    Created hld_typeahead
  6. Aug 8, 2026
    Created spend_native — A personal expense tracker app/widget
  7. Aug 19, 2026
    Created rewardhack-bench — Predict how a policy games a reward. RL environments scored by an executable-exploit verifier.
  8. Aug 20, 2026
    Created sycophancy-bench — Do models catch flawed ML results? Flawed reports with byte-identical matched controls.
  9. Aug 20, 2026
    Created silentbug-bench — ML training defects that never crash. An agentic benchmark with patch-and-rerun, unit-test verification.
  10. Aug 21, 2026
    Created scrapeverse
  11. Sep 2, 2026
    Created devops-assignments — DevOps assignments: Linux, shell scripting, networking, Git, Docker
  12. Sep 2, 2026
    Most recent push to devops-assignments

07 · Compare

github.com/
Json604 · 6dmedian coder

08 · Rubric

How this score was produced

Overall = Σ (category × weight) + gentle top-end curve

CategoryWeightScoreContrib.
Raw total65.0
Top-end curve+5.7
Final overall70.7

Tier thresholds

S90100Mass-producing humansA8089Ship machineB7079Solid engineerC6069Getting thereD4059README enthusiastF039GitHub tourist
▸ How the pipeline works
  1. 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.
  2. 02Triage.A small model reads every repo's file tree + README and picks the 20 files per repo that actually reveal how you code.
  3. 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.
  4. 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.
  5. 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.
Json604 · 70.7/100 — Rate My GitHub