Where Are Batch Processing Scripts and Prompts Located in CareerOps?

Batch processing scripts and prompts in CareerOps are located in the top-level batch/ directory, which contains self-contained Node.js ES modules and Markdown prompt templates for bulk operations.

In the santifer/career-ops repository, batch processing scripts and prompts are deliberately isolated within the batch/ directory to support large-scale job-search automation without interfering with interactive modes. This folder houses executable .mjs files and human-readable templates that handle token accounting, bulk evaluation, and tracker synchronization.

Location of Batch Processing Scripts and Prompts

The CareerOps repository organizes all bulk-operation logic under the top-level batch/ directory. This deliberate isolation ensures that batch-mode functionality operates independently of the core interactive CLI, allowing you to run bulk jobs without touching interactive modes. According to the source code analysis, every script in this folder is a self-contained Node.js ES module (.mjs) that can be executed directly.

Core Batch Processing Scripts

The batch/ directory contains several specialized modules designed for specific bulk workflows. Each script can be invoked directly using Node.js from the repository root.

aggregate-tokens.mjs

The batch/aggregate-tokens.mjs script calculates total token consumption across evaluated job offers. This module aggregates usage statistics for cost tracking and reporting:

node batch/aggregate-tokens.mjs

merge-tracker.mjs

The batch/merge-tracker.mjs module safely merges batch-generated tracker-addition TSV files into the main application tracker. As implemented in santifer/career-ops, it updates data/applications.md with new entries:

node batch/merge-tracker.mjs --add-dir batch/tracker-additions

reserve-report-num.mjs

For parallel batch processing, batch/reserve-report-num.mjs allocates a contiguous block of report numbers to prevent collisions between workers:

node batch/reserve-report-num.mjs --count 10

This outputs a reserved range such as 042-051, which parallel workers can claim for unique report identification during concurrent executions.

Batch Prompt Templates

Human-readable prompt templates reside alongside the scripts to standardize LLM interactions during bulk evaluations. The primary template, batch/batch-prompt.md, provides the instruction set for evaluating multiple job description URLs simultaneously.

You can leverage this template when running bulk evaluations:

node batch/batch-prompt.mjs < list-of-urls.txt

The separation of prompts (.md) from execution logic (.mjs) allows you to modify evaluation criteria without changing the underlying batch processing scripts, ensuring flexibility in bulk automation workflows.

Summary

• batch processing scripts and prompts reside exclusively in the top-level batch/ directory of the CareerOps repository.
• Three core scripts handle automation: aggregate-tokens.mjs for token accounting, merge-tracker.mjs for tracker synchronization, and reserve-report-num.mjs for parallel job coordination.
• All .mjs files are self-contained Node.js ES modules executable via node batch/<filename>.
• The batch-prompt.md template standardizes bulk evaluation instructions for LLM processing of multiple job descriptions.

Frequently Asked Questions

Where exactly are the batch processing scripts located in the CareerOps repository?

All batch processing scripts are located in the batch/ directory at the repository root. This folder contains Node.js ES modules like aggregate-tokens.mjs, merge-tracker.mjs, and reserve-report-num.mjs, along with prompt templates such as batch-prompt.md.

How do I execute a batch script in CareerOps?

Run any .mjs file directly with Node.js from the repository root. For example, executing node batch/aggregate-tokens.mjs aggregates token usage across evaluations without requiring additional build steps or configuration.

What is the purpose of the batch-prompt.md file?

The batch/batch-prompt.md file contains the human-readable prompt template that batch workers use to evaluate multiple job description URLs consistently. It standardizes LLM instructions for bulk processing, ensuring uniform evaluation criteria across parallel jobs.

Can I run batch operations in parallel without conflicts?

Yes. Use batch/reserve-report-num.mjs to allocate unique report number ranges for parallel workers. This prevents collisions during concurrent batch executions by reserving blocks like 042-051 for specific worker instances.

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