Where to Put Personal Customizations in Career-Ops Configuration: The Complete Guide

Place all personal customizations in config/profile.yml, modes/_profile.md, portals.yml, and specific template files while leaving system defaults untouched to ensure updates never overwrite your settings.

The santifer/career-ops repository separates user-specific data from system-wide defaults through a strict file hierarchy. By confining your edits to designated personal customization files, you protect your bespoke settings from being overwritten when pulling upstream updates.

Core Identity and Targeting Configuration

Your professional identity, compensation expectations, and job targeting logic live in centralized YAML files that drive the entire system's scoring and generation pipeline.

Profile Details in config/profile.yml

The config/profile.yml file serves as the single source of truth for your identity and job-targeting logic. All modes read this file to drive candidate scoring, work-authorization signals, and CV generation.

Store your name, contact information, target roles, professional narrative, compensation requirements, and location filters here:

candidate:
  name: Jane Doe
  email: jane@example.com
  location: Berlin, DE
target_roles:
  - Senior Backend Engineer
  - AI Product Manager
narrative:
  headline: "Building AI‑first products that scale"
  superpowers: "Distributed systems, LLM integration"
compensation:
  target_range: "120k-150k EUR"
  currency: EUR

Archetype Definitions in modes/_profile.md

The modes/_profile.md file contains the Target Roles definition table that shapes offer scores and CV language. Edit the Markdown table under the Target Roles heading to determine how opportunities are evaluated and how your resume is framed for each archetype.

| Archetype               | Thematic axes            | What they buy |
|------------------------|--------------------------|---------------|
| Senior Backend Engineer | Go, microservices, CI/CD | Scalable systems |
| AI Product Manager      | Product vision, LLMs, data pipelines | AI‑driven products |

Job Portal and Search Configuration

Control which jobs are scraped and how titles are filtered through dedicated portal configuration files.

Search Queries in portals.yml

Copy templates/portals.example.yml to portals.yml in your project root to configure keywords, search queries, and tracked companies. This file determines which postings are scraped and how titles are filtered:

title_filter:
  positive:
    - "backend engineer"
    - "machine learning"
  negative:
    - "intern"
    - "senior java"
search_queries:
  - "backend engineer site:linkedin.com/jobs"
tracked_companies:
  - "Google"
  - "Microsoft"

Visual and Content Customization

Modify the output appearance and supporting documents without altering the core logic.

CV Template Styling in templates/cv-template.html

Adjust design tokens, colors, fonts, and layout in templates/cv-template.html. The accompanying CSS and fonts/ folder control HTML/LaTeX rendering. Keep modifications ATS-friendly while personalizing the visual presentation:

<style>
  :root {
    --primary-color: hsl(187,74%,32%);
    --accent-color: hsl(270,70%,45%);
    --font-heading: "Space Grotesk", sans-serif;
    --font-body: "DM Sans", sans-serif;
  }
</style>

Negotiation Scripts in modes/_shared.md

Replace the example scripts in modes/_shared.md with your own wording for salary discussions. These provide ready-to-use text for negotiations but do not affect scoring algorithms:

### Salary Negotiation

- **Target range:** 130k–150k EUR
- **Geographic arbitrage:** Emphasize cost‑of‑living differences
- **Pushback response:** "Based on market data and my impact, I’m confident this range reflects my value."

Advanced and Optional Customization

Extend functionality through IDE-specific hooks and workflow modifications.

IDE Hooks in .claude/settings.json

Add session-start commands and integrations in .claude/settings.json for Claude Code. OpenCode users should use .opencode/commands/ or .opencode/agents/ instead:

{
  "hooks": {
    "SessionStart": [{
      "hooks": [{
        "type": "command",
        "command": "echo 'Career‑ops session started'"
      }]
]
  }
}

Workflow State Extensions in templates/states.yml

Only modify templates/states.yml when extending the tracker's state machine. This file defines canonical workflow states, with related logic in normalize-statuses.mjs and references in modes/_shared.md. Most users should never edit this file unless adding entirely new tracker states.

Summary

• Edit only config/profile.yml, modes/_profile.md, portals.yml, templates/cv-template.html, and modes/_shared.md for standard personal customizations.
• Never modify system defaults in untouched template files unless you intend to change shared behavior for every user.
• Copy example files (such as templates/portals.example.yml to portals.yml) before editing to establish your personal layer.
• Preserve update safety by keeping custom hooks in .claude/settings.json and rare state modifications in templates/states.yml.

Frequently Asked Questions

Will updating the career-ops repository overwrite my personal settings?

No, provided you edit only the designated personal customization files. The repository architecture intentionally separates user-specific data in config/profile.yml, portals.yml, and user-copied templates from system-wide defaults. Files you create or modify in these specific paths persist safely through git pulls and updates.

How do I configure job search keywords for specific portals?

Create a portals.yml file in your project root by copying templates/portals.example.yml. Edit the title_filter sections for positive and negative keyword matching, and modify search_queries to control which jobs are scraped from each portal.

Can I customize the visual appearance of generated CVs?

Yes, modify templates/cv-template.html to adjust CSS design tokens such as --primary-color, --accent-color, and font families. You can also replace fonts in the fonts/ folder. Maintain ATS-friendly standards while customizing the visual output for PDF generation.

Where should I add custom automation hooks for Claude Code?

Define custom hooks in .claude/settings.json using the SessionStart event type to run commands when your IDE session begins. These optional extensions allow integrations without affecting the core career-ops logic or other users' environments.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}