| name | career-prep |
| description | AI-powered career preparation: evaluate JDs, generate cover letters & CVs, research companies, scan jobs, practice interviews, negotiate offers. Works with the Interview Prep Portal React dashboard. |
| version | 1.0.0 |
Career Prep — AI-Powered Job Search System
Overview
Career Prep is a Hermes Agent plugin that turns the agent into a complete job search command center. It works standalone through the agent's tools, slash commands, and CLI, plus pairs with the Interview Prep Portal web dashboard for visual tracking.
Inspired by career-ops.
Quick Start
Enable the plugin
hermes plugins enable career-prep
Restart your Hermes session. The /prep slash command is now available.
One-shot via CLI
hermes career-prep evaluate "https://jobs.example.com/ai-engineer"
hermes career-prep cover "Anthropic" "AI Engineer"
hermes career-prep research "OpenAI" --role "MCP Engineer"
hermes career-prep serve
hermes career-prep status
Web Portal
hermes career-prep serve
The portal provides visual tracking for applications, learning paths, flashcards, and company research.
Tools
evaluate_jd
Evaluate a job description against your profile.
- Input: URL to job posting or pasted JD text
- Output: Structured A-F evaluation with score, match analysis, prep roadmap
- Use when: User pastes a job link, asks "is this role good for me?"
generate_cover_letter
Generate a tailored, ATS-optimized cover letter.
- Required: company name, role title
- Optional: job description (for keyword matching), angle (impact/why_them/why_me/mission)
- Use when: User needs a cover letter for a specific application
research_company
Deep research on a target company.
- Input: company name, optional role
- Output: Products, tech stack, culture, recent news, interview process, talking points
- Use when: Before an interview or deciding whether to apply
scan_jobs
Search job boards for matching roles.
- Optional: search terms, location, max results
- Use when: "Find me AI Engineer jobs in Toronto"
generate_interview_stories
Create STAR+Reflection stories for behavioral interviews.
- Focuses: leadership, conflict, failure, innovation, collaboration, technical
- Use when: Preparing for behavioral interviews
generate_negotiation_script
Generate salary negotiation talking points.
- Input: offer details (comp, company, role)
- Output: Anchor strategy, talking points, phrasing, leverage points
- Use when: User has an offer and needs to negotiate
serve_portal
Start/stop the Interview Prep Portal web dashboard.
portal_status
Quick summary of your application pipeline and study progress.
Slash Commands
| Command | Description |
|---|
/prep | Show all commands |
/prep evaluate <url> | Evaluate a job description |
/prep cover <co> <role> | Generate cover letter |
/prep research <co> | Research a company |
/prep scan | Find new job opportunities |
/prep stories | Generate interview stories |
/prep negotiate | Get a negotiation script |
/prep portal | Open web dashboard |
/prep status | Your application stats |
User Profile
Configure your profile in the Interview Prep Portal web dashboard for personalized evaluations. The default profile is intentionally blank so the tools do not leak a maintainer persona into another job seeker's materials.
Fill in:
- Identity: name, contact, location, work authorization
- Career: current title, level, industry, target roles
- Skills: core skills, growing skills, certifications
- Compensation: salary/rate targets, negotiability
- Work history: concrete outcomes, tools, scope, and story seeds
- Agent: Hermes, Claude Code, Codex, HTTP, or offline mode
Edit through Settings or Onboarding. The backend profile file is the source of truth for AI-generated materials.
Architecture
User → Hermes Agent → /prep slash → Tool dispatch → local backend
│
┌────┴────┐
│ Plugin │
├─────────┤
│ evaluate_jd → POST /api/evaluate_jd
│ cover_letter → POST /api/cover_letter
│ research_company → POST /api/research_company
│ scan_jobs → POST /api/scan_jobs
│ interview_stories → POST /api/interview_stories
│ negotiation_script → POST /api/negotiation_script
│ portal_status → GET /health + /profile
└─────────┘
│
Python backend
│
Profile + chosen AI agent
│
Web Portal (React SPA)
localhost:8766
Tips
- First evaluations won't be perfect — feed the system more context through the portal
- Score threshold: Only apply to roles scoring 4.0+/5
- The portal is your truth — all data persisted in localStorage, export as JSON
- Combine tools: evaluate a JD, then generate a cover letter, then research the company
- Use the portal for spaced repetition: 30+ flashcards, 7 learning paths, 35+ curated resources