← Back to Agent Skills
🤝

network-jobs

v2.0.4 • MIT • View on GitHub

Install

npx skills add https://github.com/hirefrank/skills --skill network-jobs

Network Jobs

Search job openings at companies where you have connections through your personal network.

Configuration

  • BASE_URL: https://jobs.hirefrank.com
  • ADVISOR_SLUG: hirefrank (default, use a different slug only if explicitly mentioned)

Fetching Data

Data is hosted at jobs.hirefrank.com. Fetch files directly:

# Fetch any file
curl -s "https://jobs.hirefrank.com/{ADVISOR_SLUG}/{file}.json"

# Examples
curl -s "https://jobs.hirefrank.com/hirefrank/advisor.json"
curl -s "https://jobs.hirefrank.com/hirefrank/manifest.json"
curl -s "https://jobs.hirefrank.com/hirefrank/engineering.json"
curl -s "https://jobs.hirefrank.com/hirefrank/product-nyc-senior.json"

Data Schema

advisor.json

{
  "name": "Frank Miller",
  "slug": "hirefrank",
  "email": "frank@example.com",
  "title": "VP Engineering",
  "company": "Acme Corp",
  "url": "https://linkedin.com/in/frankmiller"
}

manifest.json

{
  "advisorSlug": "hirefrank",
  "lastUpdated": "2026-01-09T22:30:00Z",
  "totalJobs": 430,
  "categories": {
    "product": {
      "count": 679,
      "file": "product.json",
      "byLocation": {
        "nyc": { "senior": { "count": 71, "file": "product-nyc-senior.json" }, "mid": { "count": 59, "file": "product-nyc-mid.json" } },
        "sf": { "senior": { "count": 66, "file": "product-sf-senior.json" }, "mid": { "count": 79, "file": "product-sf-mid.json" } },
        "remote": { "senior": { "count": 82, "file": "product-remote-senior.json" }, "mid": { "count": 53, "file": "product-remote-mid.json" } },
        "other": { "senior": { "count": 97, "file": "product-other-senior.json" }, "mid": { "count": 172, "file": "product-other-mid.json" } }
      }
    }
  }
}

Location buckets: nyc (New York), sf (San Francisco/Bay Area), remote, other Seniority buckets: senior (Senior, Staff, Principal, Lead, Director, VP), mid (everything else)

Category Files (e.g., engineering.json)

[
  {
    "id": 123,
    "title": "Senior Backend Engineer",
    "company": "Stripe",
    "companyDomain": "stripe.com",
    "department": "Developer Infrastructure",
    "category": "engineering",
    "location": "San Francisco, CA",
    "url": "https://stripe.com/jobs/123",
    "salary": { "min": 180000, "max": 250000 },
    "postedAt": "2025-12-15T00:00:00Z",
    "firstSeen": "2026-01-01",
    "lastSeen": "2026-01-09"
  }
]

Date fields:

  • postedAt: When the job was originally posted (from ATS). Use this for freshness display.
  • firstSeen: When we first discovered the job (fallback if postedAt is missing)
  • lastSeen: Last time the job was confirmed still open

Categories

engineering | product | design | data | ai-ml | sales | marketing | customer-success | operations | finance | people | legal | it-security | retail | other

For detailed category mapping (which user queries map to which category), see reference.md.

Procedures

First step for all patterns: Fetch advisor info:

curl -s "https://jobs.hirefrank.com/hirefrank/advisor.json"

Pattern A: "Do I have connections at [Company]?"

  1. Fetch Manifest:
    curl -s "https://jobs.hirefrank.com/hirefrank/manifest.json"
  2. Fetch Categories: Download relevant category files (or all if unsure)
  3. Search: Filter where company matches (case-insensitive)
  4. Report:
    • Found: "Yes! [X] open roles at [Company]."
    • Not found: "No current openings at [Company] in your network."

Pattern B: "Find me [Role] jobs in [Location]" (PREFERRED - use granular files)

  1. Fetch Manifest: Get category list with byLocation sub-files
  2. Map Category: "PM/product" -> product, "engineering/dev" -> engineering, etc.
  3. Map Location: "NYC/New York" -> nyc, "SF/Bay Area" -> sf, "remote" -> remote
  4. Fetch Granular File: Use {category}-{location}-{seniority}.json for fastest results
    • Example: "Senior PM jobs in NYC" -> fetch product-nyc-senior.json (26KB instead of 242KB)
    • If no seniority specified, fetch both senior and mid files for that location
  5. Report: Results sorted by lastSeen (freshest first)

Pattern B2: Broad queries - ASK FOR CLARIFICATION

If user query is too broad (e.g., "find me jobs", "what's available?"), ask:

  • "What type of role? (engineering, product, sales, etc.)"
  • "Any location preference? (NYC, SF, remote, anywhere)"
  • "Seniority level? (senior/staff or mid-level)"

This avoids fetching 200KB+ of data unnecessarily.

Pattern C: "What's new?" / "Recent opportunities"

  1. Fetch Manifest: Check lastUpdated timestamp
  2. Strategy: Fetch user's preferred category granular file, or ask for preference
  3. Sort: By firstSeen descending to show newest additions
  4. Report: Latest 10 jobs added to the network

Pattern D: "Remote [Role] jobs" or "[Location] jobs"

  1. Map Location to bucket: nyc, sf, remote, other
  2. Fetch Granular Files: {category}-{location}-senior.json and {category}-{location}-mid.json
  3. Report matches (no client-side filtering needed!)

Pattern E: "Jobs paying over $X"

  1. Fetch appropriate category
  2. Filter where salary.min >= X or salary.max >= X
  3. Report matches with salary ranges

Output Format

STRICT FORMAT REQUIRED - Do NOT summarize or paraphrase. Do NOT write prose/narrative. Output MUST follow this exact structure:

Structure:

  1. Header: "Searching via [Name] ([Title])..." or "Searching via [Name] ([Title] @ [Company])..." if company is set
  2. Data freshness: "Data as of [Mon D], [H:MM AM/PM] ([relative] ago)" - from manifest.lastUpdated
    • Calculate relative time: use "Xh ago" for <24h, "Xd ago" for ≥24h
  3. Summary line: "X [role] roles in [location]:" (simple count)
  4. Group results by company (COMPANY NAME in caps, then "- N roles")
  5. Each job on its own line with bullet • , title, salary (if available), days since posted
  6. CRITICAL: End each job line with a clickable markdown link: [↗](job_url) - do NOT show full URLs
  7. Calculate days from postedAt (or firstSeen as fallback). Display as 3d, 14d, 120d
  8. Footer: "Want help drafting an intro email to [Name] ([email])?" - MUST include advisor email

Job line format:

• [Title] – [Salary if available], [N]d [↗](url)

Example format:

Searching via Frank Harris (Executive Coach)...
Data as of Jan 9, 10:30 PM (2h ago)

8 PM roles in NYC:

JUSTWORKS - 5 roles
• Group PM, Growth – 26d [↗](https://example.com/job/1)
• Senior PM, Foundations – $150-180k, 39d [↗](https://example.com/job/2)
• Senior PM, Payroll & Tax – 28d [↗](https://example.com/job/3)
• Product Manager, Core Services – 19d [↗](https://example.com/job/4)
• Group PM, Internal Tools – 18d [↗](https://example.com/job/5)

FIGMA - 2 roles
• Product Manager, CMS – 67d [↗](https://example.com/job/6)
• Product Manager, Growth – 66d [↗](https://example.com/job/7)

BETTERMENT - 1 role
• Sr. PM, Employer Recordkeeping – 65d [↗](https://example.com/job/8)

Want help drafting an intro email to Frank (frank@example.com)?

DO NOT:

  • Write narrative summaries like "I found 93 roles including..."
  • Group jobs by description instead of listing each one
  • Omit the clickable [↗](url) links
  • Skip the advisor email in the footer
  • Add commentary between job listings

Important Notes

  • Fetch data directly from https://jobs.hirefrank.com/{advisor}/{file}.json
  • Every job in this database is at a company where the advisor has a connection
  • The department field is the company's internal team name (useful for specificity)
  • category is the normalized job function
  • Jobs are refreshed every 6 hours; check lastUpdated in manifest

Related Skills

intro-email-generator - Once you find a job you like, use this skill to craft a compelling, forwardable introduction email. Pass the advisor info (name, email) and job details. The email is addressed to the advisor so they can forward it directly to someone at the company.