$ loading_
Score agent outputs with guardrails, policy checks, injection, and PII detection.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "ai.responsibleailabs/rail-score" MCP server from askskill: Run: claude mcp add --transport http 'ai-responsibleailabs-rail-score' 'https://mcp.responsibleailabs.ai/mcp'
Use rail-score to score this agent response, check for policy violations, prompt injection signals, or PII leakage, and return the risk results for each category: {{agent_output}}A risk scoring result showing policy risk, injection risk, and PII detection findings.
Run guardrail checks on the following user input, determine whether it contains prompt injection or sensitive information issues, and provide a score: {{user_prompt}}Returns a safety score and matched issues for the input to help decide whether to block or continue.
Please batch-check this set of model outputs for policy compliance, PII risk, and injection risk, and list the items needing manual review from highest to lowest risk: {{outputs}}A risk-ranked list of detection results for manual review and quality control.
Developers can use it to score agent inputs or outputs and detect policy violations, prompt injection, and PII risks. This adds a safety check before execution or before returning results.
Before releasing new prompts, workflows, or agent capabilities, teams can use this tool for guardrail scoring. It is suitable for finding potential safety and compliance issues.
When an application handles user text or model-generated content, it can be used to detect whether PII is present. This helps reduce the risk of exposing sensitive information.
It is a Responsible AI guardrails tool for agents. Based on the description, it supports policy scoring, prompt injection detection, PII detection, and DPDP-related capabilities.
The explicitly stated capabilities include policy risk scoring, prompt injection detection, and PII detection. DPDP is also mentioned in the description, but its exact details should be checked in the source repository.
The provided materials do not include installation steps, runtime requirements, or key information. Please see the source repository for integration details and prerequisites.
Evaluate file write requests for sensitive data, compliance, and auditable risk.
Scan AI agents for tool-calling vulnerabilities and surface key security risks.
Add human approval and tamper-evident logs to risky AI agent actions.
Scan MCP servers and AI tools for risks with scoring and auto-protection.
Analyze AI agent traces to diagnose failures and recommend actionable improvements.
Detect leaked secrets, prompt injection, and PII in AI output.