Skip to content
← Tools / Skills
Skill · .md packagedungnotnull/10 trustMetrics-checked
mock-interview-coach logo

mock-interview-coach

Mock interview simulator that scores answers against STAR/CAR rubrics and builds improvement roadmaps.

Visit official site →

About

mock-interview-coach is an agent skill that simulates realistic interviews across industries and roles, then grades answers against validated rubrics — STAR/CAR, competency-based, and MECE frameworks — rather than giving vague encouragement. It produces specific feedback on structure, content, and delivery, rewrites weak answers using proven frameworks, and builds targeted practice plans, grounding its approach in structured-interview validity research from I-O psychology literature. Installation is a git clone plus a Python dependency install for its knowledge updater. At 4 stars with an MIT license it is thinly vetted, so treat the rubric quality as promising but not yet community-hardened. Useful for new graduates drilling behavioral questions, engineers prepping for loops, and career switchers learning to frame pivots credibly.

What it can do

  • Multi-industry mock interview simulation
  • STAR/CAR rubric-based answer scoring
  • Answer rewrites using proven frameworks
  • Improvement roadmaps and practice plans
  • Delivery and structure feedback
  • Knowledge updater tooling

Setup

  1. Clone the repo: git clone https://github.com/dungnotnull/mock-interview-coach-agent-skill.git
  2. Install Python dependencies for the knowledge updater: cd tools && pip install -r requirements.txt
  3. Add the skill to your Claude Code configuration.
  4. Start a session and request a mock interview for your target role and industry.

Prerequisites

Claude Code; Python 3 with pip for the knowledge updater tools; a target role/industry to simulate.

Use cases

interview-prepmock-interviewscareer-coaching

Works with

Claude Code

Trust score — 6/10

/10 trustMetrics-checked

Scored 6/10 (metrics tier, metrics-v1). Evidence: 4 GitHub stars verified via API; MIT license confirmed; STAR/CAR rubric approach documented in README; Thin community vetting; install command verbatim.

★ 4 GitHub stars

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.