requesty-guardrails

star 0

Configure enterprise-grade security filters for AI requests — PII detection, secret key scanning, and financial data masking through Requesty's guardrails. Use this skill when the user asks about protecting sensitive data in AI requests, blocking PII from reaching LLMs, detecting API keys or credentials in prompts, masking credit card numbers, GDPR compliance for AI, SOC 2 compliance, or content filtering. Also use when the user mentions guardrails, data masking, prompt security, preventing data leaks through AI, or protecting financial information in LLM requests.

requestyai By requestyai schedule Updated 6/7/2026

name: requesty-guardrails description: Configure enterprise-grade security filters for AI requests — PII detection, secret key scanning, and financial data masking through Requesty's guardrails. Use this skill when the user asks about protecting sensitive data in AI requests, blocking PII from reaching LLMs, detecting API keys or credentials in prompts, masking credit card numbers, GDPR compliance for AI, SOC 2 compliance, or content filtering. Also use when the user mentions guardrails, data masking, prompt security, preventing data leaks through AI, or protecting financial information in LLM requests.

Requesty Guardrails

Enterprise-grade security filters that automatically detect and mask sensitive information in both AI requests and responses — PII, credentials, financial data — before they reach any model.

Prerequisites

How Guardrails Work

User request → Input scanning → Mask sensitive data → AI model → Output scanning → Mask response → Clean response

Guardrails are bidirectional — they scan both what you send to the model AND what the model sends back. Sensitive data is masked before it crosses any boundary.

What Guardrails Detect

PII (Personally Identifiable Information)

Data type Example Masked as
Social Security Numbers 123-45-6789 [SSN REDACTED]
Email addresses user@company.com [EMAIL REDACTED]
Phone numbers +1-555-123-4567 [PHONE REDACTED]
Names John Smith [NAME REDACTED]
Personal identifiers Driver's license, passport numbers [PII REDACTED]

Credentials & Secrets

Data type Example Masked as
API keys sk-proj-abc123... [API KEY REDACTED]
Database credentials postgres://user:pass@... [CREDENTIAL REDACTED]
Authentication tokens Bearer eyJhbGciOi... [TOKEN REDACTED]
Service account keys JSON key files [SECRET REDACTED]

Financial Information

Data type Example Masked as
Credit card numbers 4111-1111-1111-1111 [CARD REDACTED]
CVV codes 123 (in card context) [CVV REDACTED]
Bank account numbers Account/routing numbers [BANK REDACTED]
Financial data Investment details, statements [FINANCIAL REDACTED]

Enabling Guardrails

Step-by-Step

  1. Go to https://app.requesty.ai/guardrails
  2. Toggle each guardrail type ON/OFF:
    • PII Protection — emails, phones, SSNs, names
    • Secret Keys Protection — API keys, tokens, credentials
    • PCI Protection — credit cards, CVVs, cardholder data
    • Banking Protection — account numbers, routing numbers
    • Financial Data Protection — investment details, statements
  3. Changes apply immediately across all API keys in the organization
  4. No code changes needed — guardrails are applied at the gateway level

Key Properties

Property Detail
Scope Organization-wide — applies to every API key
Direction Bidirectional — scans inputs AND outputs
Activation Instant — no restart or downtime
Code changes None — transparent at the gateway layer
Bypass No exceptions or bypass mechanisms

No Code Changes Required

Guardrails work transparently. Your existing code doesn't change:

from openai import OpenAI

client = OpenAI(
    base_url="https://router.requesty.ai/v1",
    api_key="rqy_..."
)

# Even if the user message contains PII, it's masked before reaching the model
response = client.chat.completions.create(
    model="anthropic/claude-sonnet-4-5",
    messages=[
        {"role": "user", "content": "Process this order for John Smith, card 4111-1111-1111-1111"}
    ]
)
# The model never sees the real card number or name

Compliance Coverage

Standard How guardrails help
GDPR PII detection prevents personal data from reaching AI models
PCI DSS Credit card and payment data automatically masked
SOC 2 Credential and secret detection prevents exposure
HIPAA Personal health identifiers caught by PII filters

Combine with EU routing (base_url="https://router.eu.requesty.ai/v1") for full GDPR data residency.

Decision Tree

What data are you protecting?
├── Personal data (names, emails, SSNs)?
│   └── Enable PII Protection
├── API keys, passwords, tokens?
│   └── Enable Secret Keys Protection
├── Credit cards, payment data?
│   └── Enable PCI Protection
├── Bank accounts, routing numbers?
│   └── Enable Banking Protection
├── Investment data, financial statements?
│   └── Enable Financial Data Protection
└── All of the above (enterprise)?
    └── Enable all guardrails + EU routing

Links

Install via CLI
npx skills add https://github.com/requestyai/skills --skill requesty-guardrails
Repository Details
star Stars 0
call_split Forks 0
navigation Branch main
article Path SKILL.md
More from Creator