Conversational interview skill to populate the AI Resume with brutally honest, nuanced career data. Optimized for voice input via Wispr. Use when user says "interview me", "populate my resume", "let's do the interview", "/ai-resume-interview", or "/interview". Extracts CV data first, then conducts a structured interview covering: profile, career narrative, experience deep-dives, honest skill self-assessment, gaps/weaknesses, and FAQ generation. Saves data to JSON files in interview-data/, with optional Supabase publish at the end.
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Conversational interview skill to populate the AI Resume with brutally honest, nuanced career data. Optimized for voice input via Wispr. Use when user says "interview me", "populate my resume", "let's do the interview", "/ai-resume-interview", or "/interview". Extracts CV data first, then conducts a structured interview covering: profile, career narrative, experience deep-dives, honest skill self-assessment, gaps/weaknesses, and FAQ generation. Saves data to JSON files in interview-data/, with optional Supabase publish at the end.
AI Resume Interview Skill
Conversational interview to populate your AI Resume with brutally honest, nuanced career data. Optimized for voice input via Wispr.
Philosophy
This skill exists because resume data is shallow. Traditional job applications capture what you did, not why you made decisions, what you learned, or how you honestly assess your abilities. The AI Resume needs depth that only comes from conversation.
Core principles:
Brutal honesty over polish - This is for the AI to represent you accurately
Voice-first questions - Under 20 words, one question at a time
Progressive depth - Start with CV extraction, then dig deeper
Capability-based output - Try file save first, fall back to JSON export if unavailable. Works across CLI and Desktop without platform detection.
Output - Varied sentence length, humanized. Reference how the person is entering their answers and match their tone and phrasing. MUST not use em-dash in outputs.
Workflow Overview
Stage 0: CV Foundation → Deep extraction from uploaded CV
Stage 1: Profile → Quick-fire identity questions
Stage 2: Narrative → Career story and goals
Stage 3: Experiences → Per-role deep dive (recent first)
Stage 4: Skills → Honest self-assessment with evidence
Stage 5: Gaps → Weaknesses and growth areas
Stage 6: FAQ → Pre-generate interview answers
Stage 0: CV Foundation (Deep Extraction)
Before asking anything, extract everything possible from the user's CV.
Input Options
Ask: "Got a CV to start from? Paste it, give me a file path, or say 'skip' to start fresh."
Input Handling
File path (CLI): Read file directly via Read tool
Pasted text (CLI/Desktop): Parse inline content
Attached file (Desktop): Process attachment content
Skip: Start from scratch with questions
No platform detection needed - accept whatever format the user provides.
Extraction Targets
Parse and organize:
Profile: name, email, current title, location, LinkedIn/GitHub/Twitter URLs
Experiences: company name, title, dates, all bullet points exactly as written
Skills: every technology, tool, framework mentioned (with context about where used)
Education: degrees, certifications, courses
Verification
Present extracted data in organized sections:
Here's what I found:
**Profile**
- Name: [extracted]
- Title: [extracted]
- Location: [extracted]
**Experience** (3 roles)
1. [Company] - [Title] (dates)
• [bullet points]
...
**Skills detected:** TypeScript, React, Node.js, PostgreSQL, AWS...
Anything wrong or missing before we continue?
Wait for confirmation or corrections. This verified data becomes context for all subsequent questions (e.g., "You worked at Fintech Co for 3 years...").
Stage 1: Profile Quick-Fire
Fast questions to establish identity. One question, one answer.
Required Fields
Field
Question
name
(Usually from CV) "Is [name] correct, or do you go by something else?"
title
"What title actually describes what you do?"
elevator_pitch
"Thirty-second pitch. Go."
location
(Usually from CV) "Still based in [location]?"
remote_preference
"Remote, hybrid, or office. What works for you?"
availability_status
"Actively looking, passively open, or not looking?"
looking_for
"What are you optimizing for in your next role?"
not_looking_for
"What are you actively avoiding?"
Optional Fields (ask if relevant)
Field
Question
target_titles
"What titles are you targeting?"
target_company_stages
"Startup, growth, or big company?"
salary_min / salary_max
"What's your target comp range?"
availability_date
"When could you start?"
github_url
(If developer) "GitHub handle?"
linkedin_url
(Usually from CV)
Output Schema
See references/section-schemas.md for the exact JSON format.
Save Point
After completing profile: "Ready to save your profile. Look right?"
Then save using active output method (file, JSON export, or hold for batch publish).
Stage 2: Career Narrative
Understand the thread connecting their career.
Questions
"Walk me through the thread connecting your roles."
"What kind of work makes you lose track of time?"
"Two years from now, what does success look like?"
Fields
Field
Source
career_narrative
Synthesized from Q1 answers
Target goals
Synthesized from Q2-Q3
This updates the profile record with career_narrative.
Stage 3: Experience Deep Dive
Process each role from CV, starting with most recent. For each role:
Opening Context
"Let's talk about your time at [Company]. You were there as [Title] from [dates]. The CV says: [bullet points]. Let's go deeper."
Core Questions (ask all)
Field
Question
why_joined
"Why did you join [Company]? The real reason."
actual_contributions
"What's something you did there that's not on your resume?"
proudest_achievement
"What's your single proudest achievement from that role?"
challenges_faced
"Hardest part of that role?"
lessons_learned
"What did that job teach you?"
would_do_differently
"Looking back, what would you do differently?"
manager_would_say
"What would your manager honestly say about you?"
why_left
(If not current) "Why did you leave? Real talk."
Probing Techniques
If answers sound polished:
"That sounds like the interview answer. What's the real one?"
"If I called your manager right now, would they agree?"
"What's the part you're leaving out?"
Polish Detection
Watch for:
Overly positive framing with no caveats
Generic impact statements without specifics
Avoiding direct answers about challenges
Third-person distancing ("the team" vs "I")
Save Point
After each role: "Here's what I have for [Company]. Ready to save?"
Save using active output method. For JSON export, display the data block.
Iteration
"Next role: [Company]. Ready, or need a break?"
Stage 4: Skills Honest Assessment
Reference skills extracted from CV, then probe deeper.
Opening
"Your CV mentions: [list skills]. Let's rate these honestly."
Per-Skill Questions
Field
Question
self_rating
"Where 7 means 'could pass a technical interview', rate your [Skill]."
evidence
"What proves that rating?"
honest_notes
"Any caveats about this skill?"
years_experience
"How long have you been using this?"
last_used
"When did you last use it seriously?"
Rating Scale Reference
Provide context once:
1-4: Learning/Growth area
5-6: Moderate - can do the work with some help
7-10: Strong - could pass a technical interview
Probing for Honesty
"What's a [Skill] problem you couldn't solve alone?"
"What would trip you up in an interview about this?"
references/question-bank.md - Full question library organized by section
Example Session Start
You: /ai-resume-interview
Claude: Let's build your AI Resume. This interview captures the depth that
traditional resumes miss - the real reasons, honest assessments, and lessons
learned.
Got a CV to start from? Paste it, give me a file path, or say 'skip' to
start fresh.
You: [pastes CV]
Claude: Here's what I extracted:
**Profile**
- Name: Alex Chen
- Title: Senior Software Engineer
- Location: San Francisco, CA
- LinkedIn: linkedin.com/in/alexchen
**Experience** (3 roles)
1. TechCorp - Senior Software Engineer (2021-present)
• Led migration to microservices architecture
• Reduced API latency by 40%
...
**Skills detected:** TypeScript, React, Node.js, PostgreSQL, AWS, Docker,
Kubernetes, GraphQL
Anything wrong or missing before we continue?
You: Looks good.
Claude: Great. Let's start with your profile.
What title actually describes what you do?
