sim:problem

Simulates 55-minute pair programming interviews with TypeScript templates and vitest phase gates.

Updated Feb 6, 2026
One-click install
npx skills add https://github.com/kellymears/agents --skill sim-problem-kellymears
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sim:problem
Source: https://github.com/kellymears/agents/tree/main/plugins/sim/skills/sim-problem
Command: npx skills add https://github.com/kellymears/agents --skill sim-problem-kellymears

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires typescript, vitest, and includes references (resource) components.

What problem does it solve? Practicing realistic live coding interviews is hard without a human interviewer. This Skill runs a full 55-minute pair programming simulation: it scaffolds a TypeScript project with a working-but-brittle template, delivers progressive follow-ups with growing vitest test suites, tracks time, and conducts a structured debrief with a verdict. ## Core Features & Use Cases - Scaffolded challenge projects: Generates a vitest + TypeScript project with a 40-80 line template containing subtle bugs and a base test suite where some tests pass and some fail. - Progressive follow-ups: Delivers up to three conversational follow-ups that append new test blocks and npm scripts, adapting to the candidate's actual code. - Session state and timing: Persists sessions in .session.json, resumes interrupted sessions, fires time nudges at 20/40/55 minutes, and records history in .history.json. - Structured debrief: Produces a time report, code review summary, interviewer questions, communication assessment, and a Pass/Borderline/Needs Practice verdict. - Use Case: Say "give me a mock interview" and the Skill picks a problem from its bank (event bus, CSV parser, rate limiter, buggy LRU cache, and more), scaffolds the project, and runs the full interview loop. ## Quick Start Ask the assistant to start a live coding challenge or mock pair programming interview, then say "done" when your tests pass to advance phases.

Frequently Asked Questions about sim:problem

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I practice a mock coding interview with AI?▼

Trigger the skill with phrases like "coding challenge", "mock interview", or "pair programming sim". It scaffolds a TypeScript project with vitest tests, then acts as an interviewer delivering follow-ups and a debrief over a 55-minute session.

What kinds of problems does the coding challenge simulator include?▼

The problem bank covers four categories: state management (event bus, undo/redo, reactive store), data transformation (CSV parser, JSON path, schema validator), async systems (retry with backoff, rate limiter, task queue), and code review/debugging templates with subtle planted bugs.

How do I advance to the next phase of the coding challenge?▼

Say "done", "ready", or "next" when your tests pass. The skill runs the current phase's vitest suite, and if all tests pass it appends the next follow-up's test block and updates the session state.

Does the interview simulator resume if I leave mid-session?▼

Yes. Session state persists in ~/.claude/skills/sim-problem/.session.json across conversations. Any trigger while a session is active resumes it, provided the project directory still exists.

What dependencies does the coding challenge project require?▼

The scaffolded project uses only TypeScript and vitest as devDependencies, with Node 22.14.0 pinned via .node-version. The template code itself uses standard library features only, matching real interview environments.

How is the final verdict determined in the mock interview?▼

The debrief assigns Pass, Borderline, or Needs Practice based on test results per phase, code quality, and reasoning during interviewer questions. Results are appended to .history.json for tracking across sessions.