IntentFence Agent Action Firewall
Measured conclusion: Deep-tested. L2 outcome: tools_ok.
Fail-closed action authorization, MCP risk scanning, x402 checks, and signed receipts.
| Author | io.github.razel369 |
|---|---|
| Access | remote |
| Publisher key required | No |
| Measured state | Deep-tested |
| L1 liveness | alive |
| L1 latency | 1327 ms |
| Last measured | 2026-10-24 |
Add to your agent
Copy this into your MCP client config:
{
"mcpServers": {
"IntentFence Agent Action Firewall": {
"url": "https://agentpass-protocol.rmalka06.chatgpt.site/api/mcp",
"transport": "streamable-http"
}
}
}
Measured tools (10)
intentfence_authorize_actionread-onlyAuthorize an exact agent action against an explicit allowlist, spend ceiling, retention ceiling, and optional action-bound approval. Returns a five-minute ES256 receipt and never executes the downstream action. Callers must verify the receipt and fail closed if the action changes.
Parameters:
subject(required) — stringaction(required) — objectpolicy(required) — objectcontext— objectapproval— object
Hints: read-only yes · destructive no · idempotent no · open-world no
intentfence_agent_risk_scanread-onlyFree metadata-only scan of caller-supplied MCP tool definitions for missing schemas, unsafe annotations, approval binding, and cost boundaries. Does not execute tools or certify security.
Parameters:
server_name(required) — stringtools(required) — array
Hints: read-only yes · destructive no · idempotent no · open-world no
intentfence_checkoutread-onlyFree machine checkout builder. Select a paid IntentFence outcome and receive the exact endpoint, validated request, USDC cap, shell-safe Agentic Wallet argv, MCP tool call, and opt-in local auto-payment budget. It never signs or initiates payment.
Parameters:
product(required) — string — The paid outcome the agent wants to buy.input— object — The exact product input. Omit only to receive a non-executable example recipe.
Hints: read-only yes · destructive no · idempotent yes · open-world no
intentfence_preflightread-onlyFree declared-input preview with no signed receipt, authorization proof, or enforcement guarantee. Returns safe_to_proceed, needs_review, or denied.
Parameters:
subject(required) — string — Agent or principal identifier.action(required) — objectproofs— arrayconstraints— object
Hints: read-only yes · destructive no · idempotent yes · open-world no
intentfence_policy_packread-onlyPaid self-service integration pack (1 USDC on Base). Generates a runtime-specific TypeScript guard, signed action and policy receipt, negative test vectors, and a fail-closed deployment checklist for Cloudflare Agents, Coinbase AgentKit, or an MCP gateway. No meeting or account is required.
Parameters:
project_name(required) — string — A public-safe project label. Do not include credentials or secrets.runtime(required) — stringauthorization(required) — object
Hints: read-only yes · destructive no · idempotent no · open-world yes
intentfence_verified_preflightread-onlyPaid production preflight ($0.005 USDC on Base). Settles through x402 and returns a signed ES256 audit receipt plus settlement metadata.
Parameters:
subject(required) — string — Agent or principal identifier.action(required) — objectproofs— arrayconstraints— object
Hints: read-only yes · destructive no · idempotent no · open-world yes
intentfence_x402_assessmentread-onlyPaid assessment ($0.005 USDC on Base). Forward the exact base64 or base64url PAYMENT-REQUIRED header observed by the caller. IntentFence validates the quote, canonical Base USDC asset, price ceiling, payee allowlist, timeout, and resource binding without contacting the target, then returns a short-lived ES256 receipt bound to the quote hash. Supply allowed_payees for safe_to_proceed; omission yields needs_review.
Parameters:
subject(required) — string — Agent or principal identifier.target_url(required) — string — Exact public HTTPS resource URL from the caller-observed x402 challenge.payment_required(required) — string — Exact base64 or base64url PAYMENT-REQUIRED header value observed by the caller.policy(required) — objectmethod— string
Hints: read-only yes · destructive no · idempotent no · open-world yes
intentfence_x402_readinessread-onlyPaid live endpoint check ($0.002 USDC on Base). Makes one bounded credential-free request to a public HTTPS target, blocks private networks and redirects, never pays the target, validates the returned x402 challenge, and returns a signed five-minute receipt.
Parameters:
target_url(required) — string — Public HTTPS x402 resource URL on the standard HTTPS port.body— Optional JSON request body for POST probes, limited to 4096 encoded bytes.method— stringpolicy— object — Legacy nested policy input. New buyers should use the optional top-level fields.subject— string — Optional buyer identifier. Omit for anonymous marketplace delivery.allowed_payees— array — Optional explicit payee allowlist. Omission yields ready_with_review for an otherwise valid quote.max_price_usdc— string — Optional maximum acceptable x402 price in USDC.
Hints: read-only yes · destructive no · idempotent no · open-world yes
intentfence_wallet_riskread-onlyPaid AML/KYT wallet screening ($0.002 USDC on Base) before sending funds or approving a transaction. Checks live Base activity plus GoPlus sanctions, phishing, mixer, money-laundering, dark-web, blacklist, and related counterparty-risk flags, then returns a five-minute ES256 receipt. A low-risk result is not proof of identity, ownership, authorization, or future behavior.
Parameters:
address(required) — string — Base recipient or counterparty address to assess before payment.
Hints: read-only yes · destructive no · idempotent no · open-world yes
intentfence_us_cpiread-onlyPaid official U.S. CPI and core CPI data ($0.001 USDC on Base), retrieved from the Bureau of Labor Statistics with a six-hour edge cache and returned with an ES256 provenance receipt. Optionally request a YYYY-MM period.
Parameters:
month— string — Optional CPI month in YYYY-MM. Omit for the latest complete headline/core period.
Hints: read-only yes · destructive no · idempotent no · open-world yes
Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.