# Journalistic Reasoning Prover MCP for AI Agents AI Agent Connect

> Journalistic Reasoning Prover forces your AI agent to stop making up experts and start citing real, verifiable sources. It checks for false balance and missing attributions based on the SPJ Code of Ethics, ensuring your AI-generated content stands up to professional scrutiny.

## Overview
- **Category:** productivity
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_xGkY81wXH96naljNRemI9Xxej2x1apAK6AFSdqVF/ai-agent-connect
- **Tags:** journalism, fact-checking, ethics, verification, sourcing, attribution, reporting, media, prover, reasoning, hallucination, news

## Description

Journalistic Reasoning Prover forces your AI agent to stop making up experts and start citing real, verifiable sources. AI agents often sound like reporters but they fail the first test of real journalism: verification. They'll tell you studies show a trend exists without naming a single author or journal. They might present a 3 percent fringe theory as equal to a 97 percent consensus just to look balanced. This is a hallucination, not a news report.

This MCP changes how your agent handles information. Instead of letting the AI coast on vague assertions, it forces it to act like a seasoned editor. It demands named experts with full affiliations, specific study dates, and DOI links. It stops the agent from using experts say as a shortcut. It also checks for primary source corroboration, requiring at least two independent sources before a claim can be treated as a fact.

When you use this through the Vinkius catalog, your agent starts following the SPJ Code of Ethics and Kovach & Rosenstiel standards. It identifies false balance where evidence is weighted incorrectly and flags hidden conflicts of interest. It ensures every statistic includes the who, when, where, and how. You get content that actually stands up to scrutiny rather than just sounding like a news story. It turns a generic AI output into a rigorous, verifiable report.

## Tools

### validate_journalistic_reasoning
Run a structured reflection on a story to catch fabricated sources and unverified claims. It forces the agent to provide full names, dates, and primary source links.

## Prompt Examples

**Prompt:** 
```
Write a news blurb about the new tax law.
```

**Response:** 
```
### Draft Report: New Tax Law

**Summary:** The proposed tax law aims to reduce corporate rates by 5%.

**Verification Status:** ⚠️ **ACTION REQUIRED**
* **Source Fabrication:** You mentioned 'experts agree' but did not name any specific economists or organizations.
* **Missing Attribution:** Please provide the specific date of the proposal and the primary legislative body involved.
* **Next Step:** Please provide the names of the three lead economists who reviewed this proposal to proceed with the draft.
```

**Prompt:** 
```
Present the debate on vaccines.
```

**Response:** 
```
### Vaccine Policy Overview

**Current Consensus:** 97% of global health organizations support the current vaccination schedule.

**Fringe Perspective:** Some community groups express concerns regarding specific ingredients.

**Balance Check:** ✅ **Passed**
* **Note:** This report weights the 97% consensus as the primary evidence-based position while identifying the minority view as a community-specific perspective, avoiding false equivalence.
```

**Prompt:** 
```
Summarize the crime stats for city X.
```

**Response:** 
```
### Crime Statistics Summary: City X

**Report:** Violent crime increased 40% between 2019-2023.

**Verification Status:** ✅ **Verified**
* **Source:** FBI Uniform Crime Report (2024)
* **Methodology:** Reported incidents per 100,000 residents
* **Baseline:** 2019-2023 timeframe
* **Attribution:** Obtained via FOIA request on March 15, 2024.
```

## Capabilities

### Verify expert names and affiliations
The agent must provide the full name and organization of every expert mentioned.

### Corroborate claims with two sources
Every factual assertion requires independent confirmation from at least two separate sources.

### Detect and correct false balance
The agent weights perspectives based on actual evidence rather than giving equal time to fringe views.

### Enforce full WHO, WHEN, WHERE, HOW attribution
Every quote and statistic must include the source, date, location, and method of acquisition.

### Identify financial or personal conflicts
The agent checks for and discloses any interests that could bias the reporting.

## Use Cases

### Verifying a climate report
A reporter asks the agent to summarize a UN report. The agent uses validate_journalistic_reasoning to ensure the 50 percent drop target is cited with a specific date and DOI.

### Checking crime stats
A city official wants a summary of local crime. The agent uses validate_journalistic_reasoning to find the FBI UCR source, the timeframe, and the per capita baseline.

### Auditing health trends
A content creator wants to write about a new diet. The agent uses validate_journalistic_reasoning to flag the phrase studies show as a fabrication until a specific study is named.

### Reviewing political balance
An editor wants a balanced view on a new law. The agent uses validate_journalistic_reasoning to identify if a 3 percent fringe view is being given 50 percent of the coverage.

## Benefits

- Stop source fabrication by using validate_journalistic_reasoning to demand specific authors and DOI links for every study.
- Avoid false balance with validate_journalistic_reasoning, which weights perspectives based on actual evidence levels.
- Ensure complete attribution with validate_journalistic_reasoning to capture the who, when, where, and how of every quote.
- Identify hidden biases using validate_journalistic_reasoning to check for financial or institutional conflicts of interest.
- Enforce Kovach & Rosenstiel standards with validate_journalistic_reasoning to require at least two independent sources per claim.

## How It Works

The bottom line is it turns AI-generated content into verifiable journalism.

1. Provide your AI client with a news draft or a specific claim you need verified.
2. The Connector analyzes the text against the SPJ Code of Ethics and Kovach & Rosenstiel standards.
3. You get a report identifying fabricated sources, unverified claims, and balance issues.

## Frequently Asked Questions

**Can Journalistic Reasoning Prover detect fake sources?**
Yes. It flags phrases like 'studies show' or 'experts say' as fabrications if they aren't accompanied by specific names, affiliations, and journal details.

**How does Journalistic Reasoning Prover handle false balance?**
It identifies when a report gives equal weight to a fringe theory and a scientific consensus, forcing the AI to weight the coverage based on actual evidence.

**Does Journalistic Reasoning Prover check for conflicts of interest?**
Yes. It checks for financial, institutional, and personal interests that could shape reporting and requires the AI to disclose them clearly.

**Can I use Journalistic Reasoning Prover for academic research?**
Absolutely. It's excellent for ensuring that research summaries include correct citations, DOIs, and primary source corroboration.

**Does Journalistic Reasoning Prover follow the SPJ Code of Ethics?**
Yes, it is specifically grounded in the SPJ Code of Ethics, requiring the AI to seek truth, minimize harm, and act independently.

**What happens if the AI can't find a source?**
The Connector will flag the claim as unverified. It won't let the AI proceed with a vague assertion, ensuring your final content remains accurate.

**How does the prover prevent source hallucination?**
It validates that every factual claim traces to a named, verifiable primary source or study (with authors, journal, and publication year) rather than vague references.

**What journalistic ethics standards does it enforce?**
It is grounded in the SPJ Code of Ethics and the Kovach & Rosenstiel verification standards. It forces independent verification across multiple sources and flags false equivalence and conflicts of interest.

**Can it distinguish between opinion and news reporting?**
Yes, it assesses attribution, tone, and factual grounding. It flags uncorroborated assertions that are presented as objective news, requiring clear, objective evidence for every claim.