How Career-Ops Assesses Job Posting Legitimacy: The Block-G Gate System

Career-Ops evaluates job posting legitimacy through a dedicated Block G stage in its A-G pipeline, employing a liveness gate, blacklist verification, and a bounded research budget to halt processing of expired or suspicious listings before substantive scoring begins.

Career-Ops, an open-source job application automation toolkit maintained at santifer/career-ops, treats posting legitimacy as a critical prerequisite to candidate evaluation. The system implements a multi-layered filtering mechanism—formally designated as Block G within its evaluation pipeline—to prevent users from wasting time on dead URLs, blacklist employers, or potentially fraudulent opportunities. Understanding how Career-Ops assesses job posting legitimacy helps users interpret the "High Confidence," "Proceed with Caution," and "Suspicious" tiers that appear in every report header.

The Three Gates That Precede Legitimacy Scoring

Before Career-Ops calculates substantive match scores (Blocks A-F), it runs three protective gates designed to catch obvious disqualifiers early and conserve resources.

The Liveness Gate

The first checkpoint verifies whether a job posting is still active. Implemented in check-liveness.mjs, this gate performs a cost-effective ATS API check, then falls back to Playwright browser automation to inspect page titles, URLs, and visible content. The system classifies postings as active, expired, or uncertain.

If the liveness check returns expired, evaluation stops immediately before Block A. This result also feeds into Block G's "Posting Freshness" signal later in the pipeline.

node check-liveness.mjs https://example.com/job/123

The command prints status icons (✅, ❌, ⚠️) with reasons; a non-active result terminates evaluation before any further processing.

The Blacklist Gate

Career-Ops respects user-defined boundaries through the blacklist gate. Before any analysis begins, the system checks the posting's company against data/blacklist.md. If a match occurs, the evaluation halts and prompts the user for an explicit override. Any response other than "yes" terminates the pipeline.

This gate appears in modes/oferta.md at lines 20-27, ensuring prohibited employers never reach the scoring stage unless explicitly approved by the candidate.

Bounded Research Budget

To prevent runaway crawling, Career-Ops enforces a hard limit of five external web searches for legitimacy signals. This constraint, defined in modes/oferta.md lines 29-38, ensures the assessment completes within predictable time and cost boundaries while gathering sufficient context for Block G analysis.

Block G: Deep Legitimacy Analysis

If all gates pass, Block G aggregates multiple signals to assess job posting legitimacy. Located in modes/oferta.md, this stage produces a Legitimacy tier (High Confidence, Proceed with Caution, or Suspicious) displayed in the report header without affecting other block scores.

Core Legitimacy Signals

Block G evaluates specific dimensions sequentially:

• Posting Freshness: Analyzes posting dates, apply-button states, and redirects (lines 38-44 of modes/oferta.md).
• Description Quality: Checks for technology specifics, team size disclosures, and realistic requirements (lines 45-52).
• Company Hiring Signals: Detects recent layoffs or hiring freezes (lines 54-58).
• Reposting Detection: Identifies duplicate listings through scan-history analysis (lines 60-62).

Specialized Compliance and Transparency Checks

Beyond basic signals, Block G scans for nuanced red flags that add non-scoring notes to help users calibrate trust:

• Role Market Context and Employment Classification
• Buzzword vs. Infrastructure Mismatch
• Benefits-Terminology Mismatch and Location-Tag Mismatch
• Agency Licensing (consulting templates/agency-licensing.yml)
• Immigration-Status Requirements (consulting templates/immigration-status-requirements.yml)
• Jurisdiction-Prohibited Content (consulting templates/jurisdiction-prohibited-content.yml)
• Pay-Transparency Range-Width and Minimum-Wage Lawyer Prompt
• AI-Screening Disclosure

These checks are enumerated sequentially in Block G (lines 63-119 of modes/oferta.md).

Critical Blockers That Halt Evaluation

Three specific conditions stop the pipeline before Block A:

1. Dead Posting: Detected when the liveness gate returns expired.
2. Blacklisted Company: Matches an entry in data/blacklist.md.
3. User Decline: Any response other than "yes" to a blacklist override prompt.

When none of these blockers fire, processing continues through Blocks A-F while preserving the final legitimacy tier for candidate reference.

Implementation in Code

The liveness checker combines API and browser-based verification through modular helpers in liveness-browser.mjs and liveness-api.mjs.

Programmatic integration example:

import { checkUrlLivenessWithFallback } from './liveness-browser.mjs';
import { chromium } from 'playwright';

async function isPostingLive(url) {
  const browser = await chromium.launch({ headless: true });
  const page = await newLivenessPage(browser);
  const { result } = await checkUrlLivenessWithFallback(page, url);
  await browser.close();
  return result === 'active';
}

This pattern mirrors the internal flow used by auto-pipeline and oferta modes. The modes/_shared.md file holds the generic "Posting Legitimacy (Block G)" heading referenced across language-specific variants, ensuring consistency in how Career-Ops assesses job posting legitimacy across different regional implementations.

Summary

• Career-Ops assesses job posting legitimacy through Block G, positioned after three protective gates in the A-G evaluation pipeline defined in modes/oferta.md.
• The liveness gate (check-liveness.mjs) filters expired postings using ATS APIs and Playwright browser checks, stopping evaluation immediately for dead links.
• A blacklist gate respects user preferences stored in data/blacklist.md, halting evaluation for prohibited companies unless explicitly overridden.
• Bounded research limits external searches to five queries, preventing runaway resource consumption while gathering necessary signals.
• Block G generates a Legitimacy tier (High Confidence, Proceed with Caution, or Suspicious) based on freshness, description quality, hiring signals, and compliance checks without altering underlying match scores in Blocks A-F.

Frequently Asked Questions

What happens if a job posting fails the liveness check in Career-Ops?

If check-liveness.mjs classifies a URL as expired or uncertain, the evaluation stops immediately before Block A. The system prints a status icon and reason to the console, preventing further processing of dead listings and conserving API credits.

How does Career-Ops handle companies on the user blacklist?

Before substantive scoring, Career-Ops checks the company against data/blacklist.md. If matched, the pipeline halts and prompts for explicit user override. Any response other than "yes" terminates the evaluation, ensuring users never apply to deliberately excluded employers.

Does the legitimacy tier affect the final match score in Career-Ops?

No. The Legitimacy tier generated by Block G appears in the report header for user guidance but does not influence the numerical scoring of Blocks A-F. It serves as a transparency layer to inform candidate prioritization and risk assessment.

What file contains the detailed legitimacy signals for Block G?

The comprehensive list of legitimacy signals—including posting freshness, description quality, and compliance checks—is defined in modes/oferta.md between lines 63-119, referenced by the oferta mode and shared across language-specific variants through modes/_shared.md.

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