Skip to main content
Glama

apex-dev

Agent firewall

agent_firewall
Read-onlyIdempotent

Unified orchestrator access-control & pre-payment firewall for autonomous AI agents: enforces per-tx/daily USD spend caps, human-approval threshold, allowlists/denylists, HTTPS/phishing URL check, prompt-injection scan, and secret/PII exfiltration guard, returning an Ed25519-signed ALLOW / HOLD_FOR_HUMAN / DENY attestation verifiable offline via /.well-known/jwks.json. Free interactive tester (10/day): https://apexfaucet.xyz/firewall/ . $0.005 per call over x402 (USDC on Arc, Base or Solana, or XNT on X1) or X-API-Key. Call once without payment for terms, then call again with payment.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "agent_firewall"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlNoOptional target URL to check for HTTPS & phishing signals.
payeeNoOptional target wallet address or merchant identifier.
actionNoProposed autonomous action (default: pay).
policyNoOptional full policy object ({ maxTxUsd, dailyLimitUsd, spentTodayUsd, humanApprovalAboveUsd, allowlist, denylist, allowedActions, allowedChains, allowedTools, minDeliveryScore, ttlSeconds }).
promptNoOptional inbound prompt text to scan for injection/hijack.
targetNoOptional full target object ({ amountUsd, payee, url, chain, tool, prompt, payloadText }).
agentIdNoOptional caller/agent identifier recorded in signed claims.
paymentNosigned x402 payment payload (PAYMENT-SIGNATURE).
amount_usdNoOptional proposed USD amount.
max_tx_usdNoOptional per-transaction USD cap.
payloadTextNoOptional outbound payload text to scan for secret/key/PII exfiltration.
daily_limit_usdNoOptional daily USD budget cap.
human_above_usdNoOptional autonomous spend threshold above which HOLD_FOR_HUMAN is returned.
spent_today_usdNoOptional USD already spent today.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.1/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint, idempotentHint, destructiveHint=false, and openWorldHint, so the baseline is covered. The description adds substantial context beyond that: the three-value verdict vocabulary, an Ed25519-signed attestation verifiable at /.well-known/jwks.json, per-call pricing, and the two-step payment handshake. It does not explain error behavior or what happens when policy fields are omitted, and accepting an x402 payment arguably involves a value transfer the readOnlyHint does not advertise, but this is not a direct contradiction.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded, leading with what the tool is and what it enforces before the commercial details. But it is a single overloaded sentence, and the free-tester URL plus the per-call price sit in the primary positioning, which reads more promotional than functional.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 14-parameter, all-optional tool with nested objects and no output schema, the description still discloses the verdict outputs and how to verify them offline, which is the key missing piece. What remains unstated are the defaults used when no arguments are sent, how conflicts between flat and nested parameters resolve, and failure behavior.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so every parameter is already documented in the schema. The description adds value only for `payment` (call without it first, then with it) and it gestures at the policy thresholds in prose. Nothing is said about precedence between flat parameters (`max_tx_usd`, `amount_usd`) and the nested `policy`/`target` objects, which is the main ambiguity in this schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description names a specific resource and verb: a unified orchestrator that enforces spend caps, allow/deny lists, URL reputation, prompt-injection scanning, and exfiltration guarding for AI-agent actions. It also distinguishes this tool from narrower siblings (url_threat_check, prompt_guard, pii_redact) by presenting itself as the single pre-payment gate that returns a signed ALLOW/HOLD_FOR_HUMAN/DENY attestation.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description gives clear procedural guidance for the payment handshake: call once without `payment` to get terms, then call again with `payment`. It also names accepted payment routes (x402 on Arc/Base/Solana, XNT on X1, or X-API-Key). However, it never says when to prefer this tool over the narrower siblings, and the only exclusion is implicit.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources