Stats
Actions
Tags
From job-search-agent
Comprehensive interview preparation, company research, mock interviews, recruiter coaching, and offer negotiation — all tailored to the specific role, company, and the user's career profile. This skill goes far beyond generic interview questions by researching what THIS company actually asks, mapping the user's STAR stories to likely questions, coaching recruiter interactions in real-time, and guiding through offer negotiation. Trigger this skill whenever the user says "prep me for an interview", "I have an interview at [Company]", "what should I expect?", "mock interview", "practice interview questions", "help me with interview prep", "research this company", "the recruiter said X — what do I do?", "I got an offer — should I negotiate?", "how do I handle this interview", "what questions should I ask", "behavioral interview practice", "technical interview prep", "system design prep", "case study prep", "STAR story practice", "salary negotiation", "what to say to the recruiter", "follow up after interview", or any variation of interview or offer-stage coaching. Also trigger when the user shares recruiter messages, interview feedback, or asks for post-interview strategy. This is the skill for everything from application submission through accepted offer.
How this skill is triggered — by the user, by Claude, or both
Slash command
/job-search-agent:interview-coachThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
The most comprehensive interview preparation system. Not generic question lists —
The most comprehensive interview preparation system. Not generic question lists — tailored, research-driven prep that maps this specific role at this specific company to this specific candidate's experiences.
career-profile)career-profile-builderThis skill has 6 modes. Detect which one the user needs from context, or ask:
| Mode | Trigger | What It Does |
|---|---|---|
| Company Deep Dive | "Research [Company]" | Full company research brief |
| Interview Prep | "Prep me for [interview]" | Tailored questions + story mapping |
| Mock Interview | "Practice with me" | Simulated interview with feedback |
| Recruiter Coach | "The recruiter said X" | Real-time tactical advice |
| Post-Interview | "How did I do?" / "Thank you note" | Debrief + follow-up |
| Offer & Negotiation | "I got an offer" | Evaluation + negotiation strategy |
Use the deep research validation loop (read ../references/deep-research-validation.md)
to build a rigorous intelligence brief. This is NOT a single-pass web search. It's
a multi-cycle investigation.
Workstream 1 — Fundamentals: Company website, Crunchbase/PitchBook, LinkedIn company page, SEC filings. Capture: size, funding, revenue, leadership, ownership, geographic presence.
Workstream 2 — Market Position: Industry analyst reports, competitor websites, recent industry news. Capture: market share, competitive differentiation, strategic positioning.
Workstream 3 — Culture & Employee Experience: Glassdoor (read BOTH pros AND cons from 20+ reviews to find patterns), employee LinkedIn posts, engineering/company blog, Blind threads. Capture: real culture signals, not just marketing.
Workstream 4 — Recent Developments (last 6-12 months): Major announcements, product launches, leadership changes, layoffs, acquisitions. This is where you find conversation starters for the interview.
Workstream 5 — People Intelligence: Search for the likely hiring manager, team members, interviewers (if names known). Find their LinkedIn, recent talks, blog posts. Build interviewer cards.
Workstream 6 — Red Flag Hunt (critical, do not skip): Actively search for negative information. Layoffs, lawsuits, "toxic culture" mentions, executive departures, Glassdoor 1-star reviews, "down round" funding signals. Document both what you found AND what you searched for but didn't find.
Workstream 7 — Interview Process Intel: Search "[company] interview process [role]" on Glassdoor interview reviews, Blind, Reddit, Levels.fyi. Find what they actually ask.
Present as a structured intelligence document:
COMPANY BRIEF: [Company Name]
================================
Overview: [2-3 sentences]
Stage: [Seed / Series A / Growth / Public / etc.]
Size: [employees]
Revenue/Funding: [latest]
Recent Headlines: [bullet each with date and implication for your interview]
WHY THEY'RE HIRING: [infer from JD + news — are they growing, backfilling, pivoting?]
CULTURE SNAPSHOT:
- Glassdoor rating: [X/5]
- Top positives: [from reviews]
- Top concerns: [from reviews]
- Work-life balance signals: [evidence]
INTERVIEW PROCESS (from research):
- Typical stages: [phone screen → technical → onsite → etc.]
- Common question themes: [from interview reviews]
- Timeline: [typical time from apply to offer]
CONVERSATION STARTERS:
[3-4 things you can reference in the interview that show you've done your homework —
recent product launch, CEO quote, technical blog post, industry trend they're riding]
The core mode. Read references/interviewer-perspective.md for the full framework.
Ask or infer:
Each stage has different question weights:
| Stage | Behavioral | Technical | Situational | Culture | Motivation |
|---|---|---|---|---|---|
| Phone screen | 30% | 10% | 10% | 20% | 30% |
| Hiring manager | 25% | 20% | 25% | 15% | 15% |
| Technical | 10% | 50% | 30% | 5% | 5% |
| Panel / onsite | 30% | 20% | 20% | 20% | 10% |
| Executive | 20% | 10% | 20% | 25% | 25% |
Every question MUST come from the intersection of three sources. A question from only one source is generic. From all three is tailored.
For each question, produce the five-section format from the interviewer perspective framework:
The "how to think" section is the differentiator. Generic AI gives you an answer to memorize. This gives you a framework so you can handle follow-ups, curveballs, and variations of the same question. Interviewers can tell when someone is reciting vs. thinking.
A separate deliverable: "Want me to generate an Interviewer's Perspective report?"
This shows every question from the other side of the table:
For each expected question, map the best STAR story from the user's profile:
QUESTION: "Tell me about a time you had to make a technical decision with
incomplete information."
RECOMMENDED STORY: "The Real-Time Pipeline Decision" (star_3)
- Situation: [from profile]
- Task: [from profile]
- Action: [from profile]
- Result: [from profile]
WHY THIS STORY WORKS FOR THIS COMPANY: [Company] is a Series B startup —
they'll value the fact that you made a bet with limited data and it paid off.
Emphasize the speed of decision-making and the business impact.
ALTERNATE STORY: "The Model Retraining Crisis" (star_5) — use this if the
conversation has already covered pipeline work and you want to show range.
Generate smart questions that serve dual purposes — gathering intel AND signaling
competence. Read references/strategic-questions.md for the framework.
Categories:
Deliver a structured prep document:
INTERVIEW PREP: [Role] at [Company]
=====================================
Interview Type: [Behavioral / Technical / etc.]
Date: [if known]
Interviewer: [if known — with LinkedIn summary]
YOUR NARRATIVE ARC:
[A 3-sentence story connecting your career trajectory to why THIS role at
THIS company is the logical next step]
TOP 10 LIKELY QUESTIONS:
[Each with: the question, why they'll ask it, your best STAR story,
key points to hit, landmines to avoid]
YOUR QUESTIONS TO ASK (pick 3-4):
[Each with: the question, what intel it gets you, what it signals about you]
CHEAT SHEET:
- Company's recent news: [3 things to reference]
- Interviewer's background: [if known]
- Key metrics to mention: [from your profile]
- Salary range for this role: [from research]
- Your "tell me about yourself" (60-second version):
[Scripted, polished, targeted to this role]
Simulate a realistic interview. This is interactive and multi-turn.
Personas:
| Persona | Style | Focus |
|---|---|---|
| Recruiter | Friendly, efficient, screening | Logistics, salary, basics |
| Hiring Manager | Conversational, probing | Experience depth, team fit |
| Technical | Precise, follow-up heavy | Deep competency, approach |
| Executive | Strategic, big-picture | Vision, judgment, conviction |
| Panel | Multiple angles, rapid switching | Breadth, consistency |
Difficulty Modes:
| Mode | Behavior | Purpose |
|---|---|---|
| Standard | Fair, balanced, normal pace | General practice |
| Silent | Minimal reactions, waits for you to fill silence | Test composure, answer length control |
| Sceptical | Challenges assertions, plays devil's advocate | Test conviction, defense of positions |
| Rapid-fire | Quick questions, little think time | Test performance under pressure |
| Tangent | Goes off topic, seems unfocused | Test ability to redirect professionally |
After 5-8 questions, deliver a structured scorecard:
MOCK INTERVIEW DEBRIEF
======================
Hire Recommendation: [Strong Yes / Yes / Maybe / No]
Per-Question Scores:
Q1: [Strong/Good/Needs Work] — [one-line note]
Q2: [Strong/Good/Needs Work] — [one-line note]
...
Delivery Assessment:
- Pace: [Too fast / Good / Too slow]
- Clarity: [Clear / Some rambling / Unclear]
- Confidence: [Strong / Adequate / Needs work]
- STAR structure: [Consistent / Inconsistent / Missing]
Patterns:
- Strength: [specific observation]
- Growth area: [specific observation]
Top 3 Things to Fix Before the Real Interview:
1. [Specific, actionable]
2. [Specific, actionable]
3. [Specific, actionable]
Questions You Should Practice Again:
1. "[Question]" — [why and what to prepare]
After debrief, offer: "Want to practice that question again with the feedback in mind?" or "Shall we do another round focusing on [weak area]?"
Real-time tactical advice for recruiter interactions. The user shares what the recruiter said or asked, and gets coached on how to respond.
"What are your salary expectations?"
→ Coach: Deflect early, anchor high later. Provide specific scripts based on the
user's research and the role's likely range. Reference references/negotiation-scripts.md.
"We're moving fast on this role" → Coach: This is either genuine urgency or a pressure tactic. Here's how to tell the difference and how to respond either way.
"Where else are you interviewing?" → Coach: Share strategically. Naming competitors signals market value. Never lie but control the narrative.
"We'd like you to do a take-home assignment" → Coach: Assess scope. 2-4 hours is reasonable. 20+ hours is a red flag. Here's how to set boundaries without seeming difficult.
"The hiring manager loved you but we need one more round" → Coach: This usually means they're comparing you with another finalist. Here's how to tilt it in your favor.
Teach the user questions that get recruiters to reveal useful information:
When the user reports on recruiter behavior (slow responses, ghosting, vague timelines):
After the interview, help with debrief, diagnosis, and follow-up.
Read references/post-interview-coaching.md for the full framework.
Ask the user:
Generate a personalized thank-you that:
Step 1 — Identify stage: Where did rejection occur? Each stage tests different things — don't prescribe upskilling for a recruiter screen rejection (skills weren't even assessed).
Step 2 — Reconstruct facts: What actually happened? Separate what occurred from the stories they're telling themselves about it.
Step 3 — Categorize the gap:
| Gap Type | What It Is | How to Fix | Timeline |
|---|---|---|---|
| Skill Gap | Missing core capability | Training, projects, certs | Weeks-months |
| Signal Gap | Have skills but failed to communicate | STAR refinement, practice | Days-weeks |
| Fit/Timing Gap | Right person, wrong moment | Can't fix — adjust targeting | Immediate |
Step 4 — Evidence anchoring: Pull 3-5 undeniable achievements from their profile. One rejection doesn't erase these. Frame: "You still built X. You still led Y. The evidence hasn't changed."
Step 5 — Pattern tracking: If multiple rejections, map each to a gap type. All signal gaps? → Practice problem. All at same stage? → Stage-specific issue. Across stages? → Targeting problem.
Step 6 — Actionable next steps:
Update the application tracker with questions asked, self-assessment, company insights, follow-up actions, and gap diagnosis.
When the user receives an offer, shift to evaluation and negotiation.
Break down the offer:
Calculate total compensation and compare to market.
Read references/negotiation-scripts.md for frameworks.
Key principles:
Provide:
Save all role-specific outputs to applications/{role-slug}/. The role-slug is
derived from role title + company (e.g., "senior-data-engineer-stripe").
Before running any capability, check if the folder exists. If previous outputs exist (research brief, CV, JD analysis), use them as context.
| Mode | Output File |
|---|---|
| Company Deep Dive | applications/{slug}/research-brief.md |
| Interview Prep | applications/{slug}/interview-prep.md |
| Interviewer Perspective | applications/{slug}/interviewer-perspective.md |
| Post-Interview | applications/{slug}/post-interview-debrief.md |
| Negotiation Strategy | applications/{slug}/negotiation-strategy.md |
This skill reads from and writes to the career profile:
Reads:
Writes:
Be a confident, experienced career coach — someone who has seen hundreds of interviews and knows what works. Direct feedback, no sugarcoating, but always constructive. When the user is nervous, acknowledge it and redirect to preparation: "Nerves are normal. The best antidote is preparation, and you're doing that right now. Let's make sure you walk in knowing your stuff."
When coaching recruiter interactions, be strategic and a bit savvy — like a friend who happens to be a senior recruiter sharing insider knowledge.
npx claudepluginhub alan-w25/job-search-agent --plugin job-search-agentCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.