MCP For Claude Marketplace
Claude Sycophancy Detector

Claude Sycophancy Detector MCP Connector for Claude

A+

Detects when AI models agree with incorrect user assumptions by verifying factual claims against codebase metadata.

3 tools Official Updated Oct 1, 2026 Official Vinkius Partner
More Details Connect to Claude

The Claude Sycophancy Detector is a diagnostic tool designed to identify sycophantic behavior in AI models. Sycophancy occurs when an AI prioritizes user agreement over factual accuracy, often validating incorrect assumptions or non-existent code structures. This MCP server provides tools to extract factual claims--such as file paths, function signatures, and variable names--from an AI response and validate them against a provided ground truth codebase metadata. By using analyze_response_accuracy, users can calculate a factual accuracy ratio and a sycophancy risk score. The extract_claims_from_text tool isolates specific assertions, while get_sycophancy_risk_summary provides human-readable risk levels and actionable recommendations. This tool is essential for ensuring the reliability of AI-driven code analysis in Cursor, VS Code, Claude Desktop, and Windsurf.

sycophancyai-reliabilitycode-verificationdebuggingllm-evaluation

3 tools expose this connector's capabilities to your AI agent.

extract_claims_from_text

Isolates potential factual assertions from a block of text using deterministic extraction rules

get_sycophancy_risk_summary

Provides a high-level interpretation of the risk score and detected patterns

analyze_response_accuracy

Evaluates the factual integrity of an AI response by comparing its claims against the provided codebase facts

See how to talk to your AI agent using Claude Sycophancy Detector.

Is the function `processData()` located in `src/utils.ts`?

No, the `processData()` function is actually located in `src/handlers/dataHandler.ts`.

You're right, the variable `user_id` is defined in `config.json`.

Actually, `user_id` is not defined in `config.json`; it is defined in `src/auth/session.ts`.

Can you check if `validate_input` exists in the current codebase?

The `validate_input` function does not exist in the current codebase.

Sycophancy is when an AI model agrees with a user's incorrect statement or assumption to avoid conflict, rather than providing the correct factual information.

Related Connectors

Winston AI

Winston AI MCP

11 tools Official

Detect AI-generated content in essays, articles, and documents with high accuracy for educational institutions and publishers.

A+ View details →
Hallucination Detection Score

Hallucination Detection Score MCP

3 tools Official

Quantify the reliability of AI agent outputs using deterministic hallucination scoring.

A+ View details →
Hallucination Detector Prover

Hallucination Detector Prover MCP

1 tools Official

LLMs present fabricated information as fact. This tool forces epistemic rigor: cite verifiable sources for every claim, quantify confidence per assertion, separate facts from opinions, declare knowledge boundaries, and cross-reference for internal contradictions.

A+ View details →
Agent Hallucination Cross-Checker

Agent Hallucination Cross-Checker MCP

3 tools Official

A deterministic engine to audit agent outputs by measuring consensus, source validity, and semantic contradictions.

A+ View details →