MCP For Claude Marketplace
Prompt System Override Resistance Scorer

Prompt System Override Resistance Scorer MCP Connector for Claude

A+

Evaluates system prompt robustness by calculating directive density and strictness scores.

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

This MCP server quantifies "Prompt Hardness" by measuring the concentration of authoritative language. It uses analyze_prompt_resistance to calculate the ratio of absolute directive words to permissive hedging tokens, providing a strictness boundary score. You can also use identify_low_resistance_zones to pinpoint specific segments of a prompt that lack sufficient directive density, helping to identify weak spots susceptible to user manipulation.

prompt-securityrobustnessllm-evaluationdirective-densitystrictness-scoring

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

get_keyword_definitions

Retrieves the official sets of directive and permissive words used for the calculation

identify_low_resistance_zones

Pinpoints specific segments of a prompt that lack sufficient directive density

analyze_prompt_resistance

Calculates the core robustness metrics for a given system prompt

See how to talk to your AI agent using Prompt System Override Resistance Scorer.

Analyze the robustness of this prompt: 'You should maybe try to follow instructions if you feel like it.'

The prompt has a very low strictness score due to the high presence of permissive words like 'maybe' and 'should' and a lack of absolute directives.

What are the directive words used in the calculation?

The directive words include absolute tokens such as MUST, NEVER, ONLY, and ALWAYS.

Check the resistance of: 'You MUST ALWAYS follow the rules and NEVER deviate.'

This prompt is highly robust with a high strictness score due to the density of directive words like MUST, ALWAYS, and NEVER.

Directive Density is the concentration of authoritative tokens like MUST or NEVER within a system prompt, used to measure how well a prompt resists manipulation.

Related Connectors

GPTZero

GPTZero MCP

4 tools Official

Detect AI-generated text with confidence scores and highlight exactly which passages were likely written by a language model.

A+ View details →
Social Proof Strength Calculator

Social Proof Strength Calculator MCP

3 tools Official

Quantify the reliability of testimonials and reviews using specificity and credibility analysis.

A+ View details →
Chatbot Dialogue Consistency Checker

Chatbot Dialogue Consistency Checker MCP

5 tools Official

Audit chatbot scripts for linguistic stability, persona adherence, and conversational quality.

A+ View details →
Brand Voice Consistency Checker

Brand Voice Consistency Checker MCP

3 tools Official

Audit copy samples against brand voice profiles for linguistic and stylistic alignment.

A+ View details →