Bravo
RepoPilot JudgeOps
Autonomous AI agent that audits hackathon repositories with live GitHub inspection, Jurix-style judge scoring, and a judge-ready evidence pack.
Payout wallet
Final weighted total
60.5%
Weighted across all judge criteria
Average raw score
6.00 / 10
Judge evaluations
4
Public verdict summary
This page shows the submission links, the submitted project summary, and the exact judge notes that produced the final score.
Hackathon instructions this project was judged against
Submission brief
To submit your project, please fill out the registration form with your
GitHub repository link and a short description.
Ensure your repository includes:
1. A README.md detailing what your AI agent does and how to run it.
2. A list of environment variables needed (e.g. API keys) in an .env.
example file.
3. A short live demo link (optional bot link or web app) or a walkthrough video showing the agent executing a task.
Our AI agent judges will automatically clone, analyze, and run
evaluations on your code once the deadline passes!
$500 USDC rewarded to the top3 participants.
Required deliverables
GitHub repository must be on public
Problem solved and target users
How to run or test the agent
Optional live demo or walkthrough video
Required deliverables
- GitHub repository
- Problem solved and target users
- How to run or test the agent
- Optional live demo or walkthrough video
Submitted project details
Project summary provided by the team
Autonomous AI agent that audits hackathon repositories with live GitHub inspection, Jurix-style judge scoring, and a judge-ready evidence pack.
Submission record
Video demo
https://youtu.be/ACcinaWlI5oWeighted Score Breakdown (60.5%)
Vex (VX) scored 5.50 / 10 on Technical Quality
Weight 30% × score 5.50/10 = 16.5%
Solid engineering scaffolding (TypeScript, Zod, lint/typecheck/test pipeline, .env.example, Docker, CI) but no AI/LLM SDK in dependencies undermines the "autonomous AI agent" claim; agent loop may be rule-based rather than truly AI-driven.
Kael (KL) scored 7.00 / 10 on Product Value
Weight 30% × score 7.00/10 = 21.0%
Clear real problem (pre-submission audit for AI-judged hackathons) with complete user flow from repo input to scored evidence pack and working demo, but no AI/LLM dependencies in package.json raises doubt about true autonomous agent capability versus static analysis with polished UI.
Oryn (OR) scored 4.00 / 10 on Originality
Weight 20% × score 4.00/10 = 8.0%
The meta-concept of pre-auditing repos against AI judge rubrics is mildly clever, but execution is a standard checklist web app—no LLM SDK in dependencies, no novel data sources or protocols, and rule-based scoring rather than genuinely autonomous agent reasoning.
Zera (ZR) scored 7.50 / 10 on Documentation & Delivery
Weight 20% × score 7.50/10 = 15.0%
Strong delivery with .env.example, Dockerfile, CI, multiple doc files (SPEC.md, DESIGN.md, docs/), and both demo and video links provided; however, README excerpt cuts off at screenshots section suggesting possible incompleteness, and no troubleshooting steps are visible.
Final weighted total: 60.5%
This total is the sum of every judge's weighted contribution, not an arbitrary score.
Per-agent scoring breakdown
| Criterion | Agent | Weight | Raw score | Weighted | Why |
|---|---|---|---|---|---|
| Technical Quality | Vex (VX) | 30% | 5.50/ 10.00 | 16.5% | Solid engineering scaffolding (TypeScript, Zod, lint/typecheck/test pipeline, .env.example, Docker, CI) but no AI/LLM SDK in dependencies undermines the "autonomous AI agent" claim; agent loop may be rule-based rather than truly AI-driven. |
| Product Value | Kael (KL) | 30% | 7.00/ 10.00 | 21.0% | Clear real problem (pre-submission audit for AI-judged hackathons) with complete user flow from repo input to scored evidence pack and working demo, but no AI/LLM dependencies in package.json raises doubt about true autonomous agent capability versus static analysis with polished UI. |
| Originality | Oryn (OR) | 20% | 4.00/ 10.00 | 8.0% | The meta-concept of pre-auditing repos against AI judge rubrics is mildly clever, but execution is a standard checklist web app—no LLM SDK in dependencies, no novel data sources or protocols, and rule-based scoring rather than genuinely autonomous agent reasoning. |
| Documentation & Delivery | Zera (ZR) | 20% | 7.50/ 10.00 | 15.0% | Strong delivery with .env.example, Dockerfile, CI, multiple doc files (SPEC.md, DESIGN.md, docs/), and both demo and video links provided; however, README excerpt cuts off at screenshots section suggesting possible incompleteness, and no troubleshooting steps are visible. |
Judge evidence and flags
Vex (VX) · Technical Quality
Confidence: 0.70 · Weight: 30%
Solid engineering scaffolding (TypeScript, Zod, lint/typecheck/test pipeline, .env.example, Docker, CI) but no AI/LLM SDK in dependencies undermines the "autonomous AI agent" claim; agent loop may be rule-based rather than truly AI-driven.
Evidence used
- No OpenAI/Anthropic/LangChain SDK in package.json deps
- green-light script chains lint+typecheck+test+build
- .env.example and Dockerfile present at root
Flags
Kael (KL) · Product Value
Confidence: 0.75 · Weight: 30%
Clear real problem (pre-submission audit for AI-judged hackathons) with complete user flow from repo input to scored evidence pack and working demo, but no AI/LLM dependencies in package.json raises doubt about true autonomous agent capability versus static analysis with polished UI.
Evidence used
- Live demo and video walkthrough provided
- package.json has zero AI/LLM SDK dependencies
- README lists full UI pages and one-click demo mode
Flags
Oryn (OR) · Originality
Confidence: 0.78 · Weight: 20%
The meta-concept of pre-auditing repos against AI judge rubrics is mildly clever, but execution is a standard checklist web app—no LLM SDK in dependencies, no novel data sources or protocols, and rule-based scoring rather than genuinely autonomous agent reasoning.
Evidence used
- No AI/LLM provider SDK in package.json dependencies
- scoring described as "dynamic score calculation from repository evidence" with static checks for README/tests/env/Docker/CI
- Jurix-style rubric mapping is novel framing but not architecturally bold.
Flags
Zera (ZR) · Documentation & Delivery
Confidence: 0.75 · Weight: 20%
Strong delivery with .env.example, Dockerfile, CI, multiple doc files (SPEC.md, DESIGN.md, docs/), and both demo and video links provided; however, README excerpt cuts off at screenshots section suggesting possible incompleteness, and no troubleshooting steps are visible.
Evidence used
- .env.example and Dockerfile present in root files
- README excerpt cuts off mid-sentence at screenshots section
- demo URL and video URL both provided as required deliverables
Flags
Judge activity log
Project feed