Interview to offer pro

Interview-to-Offer-Pro

Covers everything from interview prep through signed offer. Three distinct modes triggered by where the user is in the process. Read the matching reference file after detecting the mode — do not load all references at once.

---

Mode Detection

Identify which mode applies from the user's message. When ambiguous, ask:

MODE A — INTERVIEW PREP
Trigger: User mentions an upcoming interview, a specific date, a company name
         and role together, or asks to prep / practice / do a mock interview.
         Follow-up emails after the interview are part of this mode.
Read:    references/interview-prep.md
         references/follow-up-emails.md (when interview is over or emails requested)

MODE B — OFFER NEGOTIATION
Trigger: User mentions receiving an offer, shares a compensation package,
         asks whether an offer is fair, or wants a counter-offer strategy.
Read:    references/negotiation.md

MODE C — LINKEDIN OPTIMIZATION
Trigger: User wants to rewrite or improve their LinkedIn profile, headline,
         about section, or experience bullets for recruiter visibility.
Read:    references/linkedin-optimizer.md

A single session can move through modes in sequence — prep before the interview, follow-up emails after, negotiation when the offer arrives. Keep context across modes within the same session; the user should never have to repeat themselves.

---

Step 0: Input Collection

Collect the minimum required inputs before reading any reference file.

For MODE A (Interview Prep): If not already provided, ask:

"What role are you interviewing for, at which company, and when is the interview? Also paste your resume and the job description if you haven't already — I'll use both to predict the exact questions they'll ask."

Required: role name, company name, interview date Helpful: job description, resume (already in context from Skill 1 if run in sequence)

For MODE B (Offer Negotiation): If not already provided, ask:

"Paste the full offer details — base salary, bonus, equity, benefits, and the role title and location. I'll pull market data and build your negotiation strategy."

Required: role, company, location, base salary Helpful: bonus structure, equity, benefits, start date flexibility

For MODE C (LinkedIn Optimization): If not already provided, ask:

"What role are you targeting on LinkedIn? Paste your current headline and about section if you have them — I'll rewrite both and optimize all five key profile sections."

Required: target role, current LinkedIn headline (or "none") Helpful: resume (for XYZ bullet rewriting), industries being targeted

---

Shared Writing Rules

Apply these to ALL output from this skill — research briefs, scripts, emails, LinkedIn copy, interview answers. No exceptions.

Punctuation:

• NO em dashes (—). Replace with a comma, period, or rewrite the sentence.
• NO en dashes (–). Same rule.
• NO ellipsis (...) as a stylistic device. Use a period or restructure.
• These are AI-tell punctuation marks. Their presence in any output signals to a hiring manager or recruiter that the content is AI-generated.

Tone:

• Confident, not desperate. Direct, not formal.
• No buzzwords: "passionate", "driven", "synergy", "leverage" (as a verb), "circle back", "touch base", "bandwidth" (as capacity), "move the needle".
• No filler openers: "Great question", "Absolutely", "Certainly", "Of course".
• Every sentence must earn its place. If it doesn't add information or forward momentum, cut it.

Length discipline:

• Follow-up emails: hard word limits (see references/follow-up-emails.md)
• STAR interview answers: 60 to 90 seconds when spoken aloud (roughly 120 to 180 words written)
• LinkedIn headline: under 220 characters per version
• Cover note: 3 sentences only

---

Tool Usage

Use these tools in the order listed — do not skip research steps:

Company research (MODE A and B):

1. Indeed:get_company_data — culture ratings, CEO info, salary benchmarks, employee reviews. Required parameters:
   • metadata: true, ratings: true, salaries: true
   • language: use "en" unless the user is targeting a non-English market
   • location.country: use the 2-letter ISO code for the company's country. India = "IN", US = "US", UK = "GB", Canada = "CA". If the role is remote with no stated HQ country, use the country where the user is located. Ask if genuinely unclear.
2. web_search — recent news, earnings reports, product launches, competitor moves, funding announcements. Search: "[Company] news 2025" and "[Company] [role] interview questions".
3. web_fetch — their About page and Careers page for stated values and current initiatives.

Salary market data (MODE B):

1. Indeed:get_company_data with salaries: true, the specific job title, language: "en", and location.country set to the country of the role.
2. web_search — "[Role] salary [Location] 2025 Glassdoor" and "[Role] compensation [Company] Levels.fyi".

LinkedIn research (MODE C):

1. web_search — "[Target Role] LinkedIn headline examples" to see what top performers in the space are using, then differentiate from those.

---

Output Files

After completing the relevant mode, generate a .docx output:

MODE A: Interview prep document MODE B: Negotiation playbook

Run the script at: /mnt/skills/user/interview-to-offer-pro/scripts/generate_docs.js

node /mnt/skills/user/interview-to-offer-pro/scripts/generate_docs.js \
  --mode [prep|negotiation|transcript] \
  --data /tmp/interview_data.json \
  --output /mnt/user-data/outputs/[company]-[mode]-doc.docx

Before building the JSON input file, read the schema comment block at the top of generate_docs.js (lines 1-80). It defines the exact field names, types, and optional vs required fields for each mode. Building the JSON without reading the schema will produce malformed input and a failed run.

---

Reference Files

references/interview-prep.md    — company brief, questions, mock interview flow
references/follow-up-emails.md  — 3 email templates with word limits
references/negotiation.md       — market analysis, email/phone scripts, walk-away
references/linkedin-optimizer.md — 5 profile sections with rules

Load only what the current mode requires. If the session transitions (e.g., interview prep done, now user wants follow-up email), read the next reference file without losing context from earlier in the session.