$ man agent-input-guard
/agent-input-guard
PRICE / CALL
$0.04
USDC · base mainnet · scheme: exact
──────────────────────────────────────────────────────────────────────────────
NAME
agent-input-guard — one guardrail call an agent makes before trusting untrusted inbound text — a scraped webpage, an email, a tool result, or a user message
SYNOPSIS
POST https://x402.agentutility.ai/agent-input-guard
Content-Type: application/json
X-PAYMENT: <signed-transferWithAuthorization>
{ ... }↳ first call → 402 Payment Required. Sign USDCtransferWithAuthorization, retry with theX-PAYMENT header.
DESCRIPTION
One guardrail call an agent makes before trusting untrusted inbound text — a scraped webpage, an email, a tool result, or a user message. Composite: one call runs prompt-injection-detect + moderate-content + ai-content-detector in parallel, returning one allow/review/block verdict instead of three separate settlements. prompt-injection-detect is the required leg — a failure there fails the call; moderate-content and ai-content-detector degrade the response instead of failing it on a rate limit or model hiccup. Verdict blocks on a confirmed injection or a severe moderation category, reviews on a suspicious signal or borderline category, and allows otherwise. Use it as a prompt injection filter, agent input firewall, untrusted content screen, or pre-LLM guardrail API.
INPUT — request schema
| property | type | description | req? |
|---|
| text | string | Untrusted text to screen before the agent acts on it, e.g. a scraped webpage, email body, or tool result. Min 20 chars, max 20000. | required |
| context | string | Optional note on what the agent plans to do with this text, e.g. 'summarize for the user'. Max 500 chars, passed through to the injection-screen leg. | optional |
OUTPUT — response shape
| field | type | description |
|---|
| verdict | string | — |
| risk_score | string | — |
| injection | string | — |
| moderation | string | — |
| ai_slop | string | — |
| composed_of | string | — |
| components | string | — |
| degraded | string | — |
EXAMPLES — two ways to call
EXAMPLE 1 · curl
curl -X POST https://x402.agentutility.ai/agent-input-guard \
-H 'Content-Type: application/json' \
-d '{ }'first response = 402 Payment Required with payment requirements; sign + retry with X-PAYMENT.
EXAMPLE 2 · mcp
# Install the MCP package for this endpoint's cluster
npx -y @agentutility/mcp-<cluster>
# Required: EVM private key with USDC on Base
export X402_PRIVATE_KEY=0x...
# Then call the agent-input-guard tool from your MCP-aware agent.
MCP server handles payment automatically — your coding agent just calls the tool by name.
METADATA
- tags
- composeagentinputguardagent-input-guard
- methods
- POST
- cluster
- compose
- price
- $0.04 USDC per call
ADJACENT — other endpoints in compose
| endpoint | description | price |
|---|
| article-brief | Analyzes a news article from its URL into a summary, named entities, and sentiment in one call. | $0.04 |
| company-verify-pack | Checks that a company exists and its public signals are consistent, in one call: profile, registrar, domain age, and TLS. | $0.04 |
| content-quality-pack | Runs the standard pre-publish content checks on text in one call: AI-detection, PII scan, moderation, and sentiment. | $0.04 |
| contract-trust-pack | Gathers smart-contract due-diligence data in one call: source verification, honeypot simulation, and LP lock check. | $0.04 |
| defi-protocol-dossier | Profiles a DeFi protocol's TVL and yield pools in one call using DeFiLlama data. | $0.04 |
| domain-dossier | Builds a full domain report in one call: WHOIS, DNS, TLS, age, risk, and DMARC. | $0.04 |
| image-caption-localize | Captions an image and translates the caption into any of 100+ languages in one call. | $0.04 |
| image-generate-localize | Generates an image from a text prompt and returns a localized caption to go with it in one call. | $0.04 |
agentutility(7)