Skip to content
View mnkprs's full-sized avatar
2 followers · 4 following

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: Quickdraw

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: Quickdraw

Block or report mnkprs

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mnkprs/README.md
Manos Kaparos — Senior Software Engineer in Athens. Full-stack products, open source, and agent evaluation.

Selected work   /   Open source   /   Expertise

I'm Manos, a senior software engineer based in Athens, Greece. I build full-stack products and tools for evaluating coding agents. My work spans Angular and React interfaces, Node.js backends but I am open in any challenging opportunities regarding latest technologies.

Selected work

fairtask ↗

Can a coding task be graded fairly?

An agent pipeline that screens SWE-bench-style tasks for unclear requirements and overly restrictive tests. It checks cited evidence against the source and compares verdicts with public human annotations.

Includes reproducible evaluations, readable agent traces, and installable agent skills.

TypeScript · Claude Agent SDK · Evaluation design

Open source

My latest upstream pull requests, newest first, with live status badges:

Project Contribution Status
Spartan Add browser regression coverage for Escape dismissal of nested popovers with tooltips. PR 1738 status
ng-select Initialize selection state for asynchronously loaded options so accessibility attributes are present on first render. PR 2883 status
TanStack Query Untrack initial client and options reads in Solid Query hooks to prevent accidental reactive dependencies. PR 11456 status
Mantine Ensure each Escape keypress closes only the top modal or drawer in a stack. PR 9188 status
ngx-bootstrap Restore AM/PM button styling in the timepicker with a class supported by Bootstrap 4 and 5. PR 6831 status

Explore my upstream pull requests →

Expertise

Area What I work with
Frontend & mobile TypeScript, Angular, React, Next.js, Capacitor, component behavior and state management
Backend & data Node.js, NestJS, PostgreSQL / Supabase, Redis, APIs, payment flows and concurrency
Agents & evaluations Agent orchestration, SWE-bench-style task screening, evidence verification, error analysis and reproducible experiments
Testing & delivery Playwright, Vitest, regression testing, GitHub Actions, mobile release pipelines
Smart contracts Solidity, Foundry, Base, contract integration and transaction receipts

Athens, Greece   ·   Explore my repositories

Pinned Loading

  1. fairtask fairtask Public

    Is this SWE-bench-style task fair to grade an agent on? Agentic screening with verified evidence, measured against human annotations. Also an agent skill: /fairtask

    TypeScript