How to Customize Archetypes and Negotiation Scripts in career‑ops

To customize archetypes and negotiation scripts in career‑ops, edit modes/_profile.md for your personal content and optionally config/profile.yml for CLI integration—never modify the system files in modes/_shared.md.

Customizing archetypes and negotiation scripts is the core personalization workflow in career‑ops, an open-source career management tool by santifer. The repository uses a layered architecture that separates immutable system logic from user-specific content, allowing you to tailor role-matching and offer negotiation without risking your changes on future updates.

Understanding the Layered Architecture

career‑ops loads content in two distinct layers. Understanding this precedence rule is essential to customizing archetypes and negotiation scripts correctly.

Layer	Purpose	Where to edit
System Layer	Default scoring rules, market vocabulary, and baseline logic.	modes/_shared.md — do not edit for personalization
User Layer	Your target roles, archetypes, narrative framing, proof points, and negotiation scripts.	modes/_profile.md (from modes/_profile.template.md) and config/profile.yml

The loading sequence is hardcoded: _shared.md is parsed first, then _profile.md is overlaid. Because your user-layer file loads last, your overrides always take precedence. This is explicitly documented in the template header at modes/_profile.template.md#L3-L11:

"This file is loaded after _shared.md, so anything you define here wins."

This design means you can safely pull repository updates without merge conflicts in your personalized content.

Customizing Archetypes in career‑ops

Archetypes define the roles you are targeting, the thematic axes that characterize each, and the value propositions that resonate with hiring managers for those positions.

Editing the Archetype Table

In modes/_profile.md, locate the Your Target Roles section. This Markdown table is where you enumerate your archetypes:

## Your Target Roles

| Archetype | Thematic axes | What they buy |
|-----------|---------------|---------------|
| **Data Engineer** | Pipelines, warehousing, GDPR | Companies that need scalable data platforms |
| **Machine Learning Engineer** | Model ops, feature stores, monitoring | Teams building production AI |
| **AI Product Manager** | Roadmapping, stakeholder alignment, metrics | Product orgs looking for AI‑driven products |

The system parses this table during evaluation mode to compute fit scores against job postings. Each row you add, remove, or reorder directly influences how modes/oferta.md scores opportunities.

Syncing Archetypes for CLI Filtering

To enable the CLI to filter and prioritize scans based on your archetypes, mirror your archetypes in config/profile.yml:

target_roles:
  archetypes:
    - name: "Data Engineer"
      level: "Senior"
      fit: "primary"
    - name: "Machine Learning Engineer"
      level: "Staff"
      fit: "secondary"
    - name: "AI Product Manager"
      level: "Senior"
      fit: "adjacent"

According to the career‑ops source code, node doctor.mjs and other entry points read config/profile.yml for raw structured data before applying your Markdown overrides. The target_roles.archetypes list drives downstream tooling behavior, including scan filtering and report prioritization.

Customizing Negotiation Scripts in career‑ops

Negotiation scripts are pre-written language blocks that the system injects into generated reports, offer-prep documents, and email drafts. These live in the same user-layer file as your archetypes.

Editing Negotiation Script Blocks

In modes/_profile.md, scroll to the Your Negotiation Scripts section. This section contains free-form Markdown that the engine extracts and inserts into the "Negotiation" block of evaluation reports.

## Your Negotiation Scripts

**Salary expectations:**
> "Based on market data for this role, I'm targeting $120K–$150K total compensation. I'm open to structuring the package to include equity or bonus."

**Geographic discount pushback:**
> "My performance is location‑agnostic; the outcomes I've delivered are remote‑first. I prefer a remote arrangement with occasional onsite syncs."

**When offered below target:**
> "I appreciate the offer. My research indicates similar roles at $130K–$150K. Could we explore a higher base or additional equity?"

Because these are plain Markdown sections, you can:

• Add new script categories (e.g., "Relocation assistance," "Signing bonus")
• Reorder existing blocks by priority
• Delete boilerplate you do not need

The evaluation engine in modes/oferta.md consumes these sections verbatim when building your personalized negotiation brief.

Step-by-Step Setup Workflow

Follow this sequence to activate your customizations for the first time:

1. Copy the profile template

   cp modes/_profile.template.md modes/_profile.md

2. Edit modes/_profile.md — Replace placeholder archetypes and negotiation blocks with your own content using the snippet patterns above.

3. Optionally configure config/profile.yml — Copy the example and fill in your details:

   cp config/profile.example.yml config/profile.yml

   Edit target_roles.archetypes to match your _profile.md table for CLI integration.

4. Verify your customization — Run a command and inspect the output:

   codex exec "career‑ops oferta <job-url>"

   Your archetype table and negotiation scripts will appear in the generated report's evaluation and negotiation sections.

How the System Processes Your Customizations

The architectural flow is implemented across three phases, as defined in the career‑ops source code:

1. Load profile — Entry points like node doctor.mjs ingest config/profile.yml for core personal data (name, location, compensation ranges).

2. Read overrides — The mode loader parses modes/_shared.md for system defaults, then immediately overlays modes/_profile.md.

3. Evaluation — When modes/oferta.md processes a job posting, it pulls your archetype table to compute fit scores and inserts your negotiation scripts into the report's final "Negotiation" block.

This data contract is enforced by the merge logic that handles the two layers. No code changes are required to personalize archetypes and negotiation scripts—only edits to the designated user-layer files.

Key Files Reference

File	Purpose
[modes/_profile.template.md](https://github.com/santifer/career-ops/blob/main/modes/_profile.template.md)	Starter template with annotated sections for archetypes, framing, and negotiation scripts.
[config/profile.example.yml](https://github.com/santifer/career-ops/blob/main/config/profile.example.yml)	Schema for structured personal data that the CLI reads before Markdown overrides.
modes/_shared.md	System defaults and scoring logic—read-only for users.
modes/oferta.md	Evaluation mode that consumes your archetypes and scripts to generate reports.

Summary

• Edit modes/_profile.md to customize archetypes and negotiation scripts—this is the sole supported user-layer file for personal content.
• Optionally edit config/profile.yml to enable CLI filtering and tooling integration for your archetypes.
• Never modify modes/_shared.md; it contains system defaults that are overwritten on repository updates.
• The overlay loading order (_shared.md first, _profile.md second) guarantees your customizations always apply.
• Changes take effect immediately on the next report generation—no rebuild or restart required.

Frequently Asked Questions

What happens if I edit modes/_shared.md instead of modes/_profile.md?

Your changes will work temporarily, but will be lost when you pull repository updates. The _shared.md file is tracked in version control and overwritten on every sync. The _profile.md file is gitignored by convention, ensuring your personalizations persist across updates.

How does the system know which archetype fits a job posting?

The evaluation engine in modes/oferta.md parses your archetype table from _profile.md and matches keywords in the job description against your "Thematic axes" column. It produces a fit score based on this intersection, weighted by your config/profile.yml fit declarations (primary, secondary, adjacent).

Can I use multiple _profile.md files for different career pivots?

The current architecture supports a single _profile.md per workspace. For distinct career pivots, maintain separate directories or branches with isolated modes/_profile.md files, then symlink or copy the appropriate version before running commands.

Do I need to restart the CLI after editing my profile files?

No. Both modes/_profile.md and config/profile.yml are read fresh on every command execution. Changes are immediate—no daemon, cache invalidation, or restart is required.

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