io.github.lazymac2x/ai-guardrails

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Cloudflare Workers MCP server: ai-guardrails

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "ai-guardrails": {
      "url": "https://api.lazy-mac.com/ai-guardrails/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (5)

• detect_pii

  Detect personally identifiable information (email, phone, SSN, credit card, IP, URL, API key) in a text string.

  Parameters:

  • text (required) — string — Text to scan for PII

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• redact_pii

  Redact PII from text. mode=mask replaces with [REDACTED]; mode=tag uses typed placeholders like [EMAIL].

  Parameters:

  • text (required) — string
  • mode — string

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• check_safety

  Check text for harmful content using weighted keyword detection. Returns a 0–1 score and verdict (safe/caution/unsafe).

  Parameters:

  • text (required) — string

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• check_hallucination

  Heuristic scoring of LLM hedge phrases that correlate with hallucinated outputs. Returns score 0–1 and risk level.

  Parameters:

  • text (required) — string

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• full_check

  Run pii, safety, and/or hallucination checks in one call. Default: pii + safety. Specify checks=["pii","safety","hallucination"] to run all.

  Parameters:

  • text (required) — string
  • checks — array

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

Repository

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.