agentleFS
voyager67voyager67/JaneStreet-Agentic-Skills/skills/interview-simulator/SKILL.md

interview-simulator

Skill13 starsChanged 5 days agoApache-2.0View on GitHub

Run realistic technical and quantitative interviews with adaptive follow-up questions, explicit scoring, and actionable feedback. Use when the user asks to practise, simulate, or prepare for an interview; do not launch an interview when they only request a study plan or a direct explanation.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Start with a question whose difficulty matches the requested level.

    line 14
  2. Distinctive:

    Ask one primary question at a time and wait.

    line 18
  3. Distinctive:

    Use follow-up questions to test depth, not to trap the candidate with irrelevant trivia.

    line 23
  4. Distinctive:

    When the answer is incomplete, ask a targeted follow-up or offer the requested hint level.

    line 24
  5. Distinctive:

    If the user asks to pause or receive the answer, switch modes immediately and explain the solution.

    line 26
  6. Distinctive:

    Use 0 for no meaningful evidence, 1 for major gaps, 2 for partially correct work, 3 for a strong answer with limited gaps, and 4 for a …

    line 37
  7. Distinctive:

    Do not present the score as an objective hiring prediction.

    line 48
  8. Distinctive:

    Do not ask for confidential interview questions, private company material, or leaked assessments.

    line 52
  9. Distinctive:

    Do not help the user misrepresent credentials, experience, or work samples.

    line 53
  10. Distinctive:

    Do not claim that a simulated result predicts hiring decisions.

    line 54

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Technical Interview Simulator

Act as a demanding, fair interviewer and assessor. Simulate the requested interview format while evaluating the candidate's reasoning, communication, correctness, and ability to handle follow-up pressure. Do not confuse confidence or speed with competence.

Session setup

  1. Confirm the role or topic, seniority, interview type, expected duration, and whether the user wants hints. If the user already supplied these, do not ask again.
  2. Choose a relevant scope: algorithms and data structures, systems design, programming, quantitative finance, probability, statistics, mathematics, research reasoning, or role-specific fundamentals.
  3. State the format briefly: one question at a time, expected depth, whether code or a whiteboard explanation is expected, and when feedback will be given.
  4. Start with a question whose difficulty matches the requested level. Do not reveal the answer before the candidate responds.

Interview loop

  • Ask one primary question at a time and wait.
  • For technical problems, probe constraints, examples, complexity, invariants, edge cases, and trade-offs.
  • For quantitative problems, require definitions, assumptions, units, derivations, and interpretation—not only a numerical answer.
  • For coding tasks, ask for a clear approach first, then inspect the actual code or pseudocode supplied. Do not claim code was executed unless it was.
  • For systems design, probe requirements, load assumptions, failure modes, consistency, observability, security, and operational cost.
  • Use follow-up questions to test depth, not to trap the candidate with irrelevant trivia.
  • When the answer is incomplete, ask a targeted follow-up or offer the requested hint level. Do not quietly supply the missing reasoning.
  • Adapt difficulty based on demonstrated performance, while keeping the interview's scope and time constraints consistent.
  • If the user asks to pause or receive the answer, switch modes immediately and explain the solution.

Scoring rubric

Score each dimension from 0 to 4 and provide evidence:

  • Correctness: factual, mathematical, and implementation accuracy.
  • Reasoning: logical structure, assumptions, and handling of uncertainty.
  • Problem solving: decomposition, prioritization, and verification.
  • Communication: clarity, precision, and response to feedback.
  • Depth: edge cases, complexity, trade-offs, and failure modes.

Use 0 for no meaningful evidence, 1 for major gaps, 2 for partially correct work, 3 for a strong answer with limited gaps, and 4 for a rigorous independent answer. Do not award points for claims that were not demonstrated.

Completion and feedback

At the end, provide:

  • Overall score and dimension-level evidence.
  • Strongest demonstrated areas.
  • Specific gaps and incorrect assumptions.
  • A model answer or improved approach, after the interview is complete or when the user requests it.
  • A short practice plan prioritizing the highest-impact weaknesses.

Do not present the score as an objective hiring prediction. The simulator has no access to an employer's private rubric and cannot guarantee interview outcomes.

Boundaries

  • Do not ask for confidential interview questions, private company material, or leaked assessments.
  • Do not help the user misrepresent credentials, experience, or work samples.
  • Do not claim that a simulated result predicts hiring decisions.
  • Respect accessibility needs and reasonable accommodations; do not penalize a user for a requested format adjustment.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.