I

InsumerAPI MCP

Reviews the InsumerAPI tools that can spend money or credits or change merchant state: verification and trust calls that use credits or pay

Reviews the InsumerAPI tools that can spend money or credits or change merchant state: verification and trust calls that use credits or pay per call in USDC, payment registrations, API key and merchant creation, merchant configuration, domain verification and the public listing. Read-only lookups keep Guard's usual handling. Off until you turn it on.

mcp.insumer · v1.0.0 · Specialized tools

By Community contributorExternal · opt-inMCP server
Protection rules
0
Permission checks
1
Mapped commands
1

Overview

What this coverage does

Reviews the InsumerAPI tools that can spend money or credits or change merchant state: verification and trust calls that use credits or pay per call in USDC, payment registrations, API key and merchant creation, merchant configuration, domain verification and the public listing. Read-only lookups keep Guard's usual handling. Off until you turn it on.

Key facts

Catalog ID
mcp.insumer
Version
v1.0.0
Kind
MCP server
Category
Specialized tools
Protection rules
0
Permission checks
1
Mapped commands
1
Maintainer
Community contributor(community, provenance recorded)
Source commit
cf0c1d3
Protection model
External · opt-in
External · opt-in

External coverage. It must be enabled through Guard controls; this page does not activate it.

Stated limits

  • Coverage is limited to the reviewed operations and the surrounding Guard policy.
  • A maintainer profile is not a security certification or an official upstream endorsement.

Command mapping

Every command this extension recognizes, with the catalog default floor Guard applies before workspace policy. Search the table, then open an operation for safe variants and the permission ID.

No per-command matchers are published for this listing. Coverage is described by the reviewed rule count, permission checks, and the surrounding Guard policy.

Tool state mapping

ToolGuard state
insumer_attestReview
insumer_wallet_trustReview
insumer_batch_wallet_trustReview
insumer_verifyReview
insumer_acp_discountReview
insumer_ucp_discountReview
insumer_buy_keyReview
insumer_buy_creditsReview
insumer_buy_merchant_creditsReview
insumer_confirm_paymentReview
insumer_setupReview
insumer_create_merchantReview
insumer_configure_tokensReview
insumer_configure_nftsReview
insumer_configure_settingsReview
insumer_publish_directoryReview
insumer_request_domain_verificationReview
insumer_verify_domainReview
otherInherit policy

Runtime identity

Catalog ID
mcp.insumer
Guard enforcement ID
command.mcp-insumer
Source path
contributions/mcp-servers/mcp.insumer.json
Contribution digest
sha256…60b01d

Mapped executables

npx

Action classes

mcp insumer tool

Technical profile

Independent technical profile

Published revision 1
Marketing distribution: off.This profile has no audience, channel, or promotion controls.
Profile content digest
sha256:092cd63fd09afd8e906a04e4b5b5400689f2036834c1727854b8c81c45214c2e
Profile source SHA
5e15ac6a9d15894463edd3854ae8f45e3af7be5f
Source version
1.0.0
Native source path
contributions/mcp-servers/mcp.insumer.json
Native extension ID
mcp.insumer
Runtime extension ID
command.mcp-insumer
Contribution digest
sha256:7974d6ea1f12467caaffc733eb854a14b58966172c7610378d5f818aa660b01d
Source contract fingerprint
sha256:32efc9a91da0fc7efd6d1299c914ef58d6ca995318db2db73bdcaf45455fc25d

Native contribution

Supported by reviewed source projection

Source version

Supported by reviewed source projection

Runtime capability

Supported by reviewed source projection

Permissions

Supported by reviewed source projection

The reviewed catalog declares 1 extension permission(s).

  • insumer_attest: review
  • insumer_wallet_trust: review
  • insumer_batch_wallet_trust: review
  • insumer_verify: review
  • insumer_acp_discount: review
  • insumer_ucp_discount: review
  • insumer_buy_key: review
  • insumer_buy_credits: review
  • insumer_buy_merchant_credits: review
  • insumer_confirm_payment: review
  • insumer_setup: review
  • insumer_create_merchant: review
  • insumer_configure_tokens: review
  • insumer_configure_nfts: review
  • insumer_configure_settings: review
  • insumer_publish_directory: review
  • insumer_request_domain_verification: review
  • insumer_verify_domain: review
  • other: inherit

Harness support

Unknown / self-reported

Two things have to line up: a client that runs the InsumerAPI MCP server, and a harness Guard supports. The server's README documents Claude Desktop, Cursor and Windsurf, and it runs under any MCP-compatible client. Guard's harness list (https://hol.org/guard/harnesses) includes Cursor and Claude Code among others; Claude Desktop and Windsurf are not on it. So the setups to expect this coverage in are Cursor and Claude Code. What we tried, on 2026-10-10: HOL Guard 3.42.0 on macOS with Claude Code 2.1.281. After hol-guard install claude-code, Guard's hooks stopped Claude Code tool calls before they ran and returned a review decision, so interception in Claude Code is confirmed. What that trial did not reach: enabling this extension and watching Guard stop insumer_attest. Enabling needs Guard's local service and a local-admin confirmation, and the trial ended before that step. What else has been checked: the catalog entry passes the repository's contribution and extension trust tests. Cursor has not been tried.

  • Claude Code on macOS: tried on 2026-10-10 with Guard 3.42.0 and Claude Code 2.1.281. Guard intercepted tool calls before they ran. This extension's review of insumer_attest was not reached in that trial.
  • Not covered: clients that reach the hosted endpoint at api.insumermodel.com/mcp by URL. That path does not go through npx, so this entry does not apply to it.
  • Claude Desktop and Windsurf run the server but are not on Guard's harness list, so this coverage does not apply there.
  • Cursor: on Guard's harness list and runs the server over npx. Not tried.

Prerequisites

Unknown / self-reported

Contributor-supplied prerequisite text is not independently verified. Four things, all ordinary. Node.js with npx, since the server is launched as npx -y mcp-server-insumer. A Guard-supported MCP client that can run a local server: Cursor or Claude Code. A way for the server to pay for metered calls: either an InsumerAPI key in INSUMER_API_KEY (free, created with an email address, 10 free verifications plus 100 requests a day) or, for x402 pay-per-call, a throwaway Base wallet holding a few dollars of USDC in INSUMER_PAYMENT_KEY; never a wallet with meaningful funds. And HOL Guard installed, with local-admin rights in your workspace, because this extension is opt-in and only a local admin can enable it.

  • Node.js with npx: the server launches as npx -y mcp-server-insumer.
  • A Guard-supported MCP client that runs local servers: Cursor or Claude Code.
  • INSUMER_API_KEY (free, created with an email address at POST https://api.insumermodel.com/v1/keys/create) or, for x402 pay-per-call, INSUMER_PAYMENT_KEY set to a throwaway Base wallet holding a few dollars of USDC.
  • HOL Guard installed, and local-admin rights in your Guard workspace to enable this opt-in extension.
  • Know the scope before you install: in our trial on macOS (Guard 3.42.0), installing Guard for Claude Code added its hooks to the user-level Claude Code settings, so Guard reviews every Claude Code session on the machine, not one project.
  • Just want to try InsumerAPI first, with no key? Connect the hosted endpoint at https://api.insumermodel.com/mcp in any client that takes an MCP server URL. It needs no key and is free within a shared daily allowance, with ten of the tools including insumer_attest and insumer_wallet_trust. This Guard coverage applies to the local npx server only, not to the hosted endpoint.

Expected outcomes

Unknown / self-reported

Contributor-supplied expected-outcome text is not independently verified. Three things, in order. The signing-keys lookup returns the published JWKS with no Guard prompt. The insumer_attest call stops at Guard's review prompt before anything is sent; nothing is spent while it waits. After approval the call completes and the agent receives a signed attestation: an id starting ATST-, pass true or false, one result per condition with the block it was read at, an ES256 signature (kid insumer-attest-v2) and a post-quantum ML-DSA-65 signature (pqKid insumer-attest-pq1), and an expiresAt 30 minutes out. The key's free verifications drop by one. Guard's activity view shows one Review decision for command.mcp-insumer with the approval recorded. If the attest call went through with no prompt, the extension is not enabled; go back to Activation.

  • Signing-keys lookup: returns the JWKS, no Guard prompt.
  • insumer_attest: Guard review prompt first; after approval, a signed attestation (ATST- id, pass true or false, per-condition result with block number, ES256 plus ML-DSA-65 signatures, expiresAt 30 minutes out).
  • Guard activity view: one Review decision for command.mcp-insumer, approved. No prompt at all means the extension is not enabled yet.

Limitations

Supported by reviewed source projection

Limitations are copied from the reviewed catalog facts. Contributor-supplied limitation notes are not independently verified. Guard reviews whether a covered call may proceed. It does not inspect the price, the conditions or the wallet in the call, and it does not cap spending; the server's own cap (INSUMER_MAX_PAYMENT_USDC for x402 pay-per-call) and InsumerAPI's credit balance do that. Only the 18 tool names in the catalog entry are reviewed. The read-only lookups, and any tool added in a later server release, inherit Guard's usual handling. The entry maps the npx launch only; the hosted endpoint at api.insumermodel.com/mcp, reached by URL, is outside it.

  • Coverage is limited to the reviewed operations and the surrounding Guard policy.
  • A maintainer profile is not a security certification or an official upstream endorsement.
  • Guard reviews the call, not the price or the conditions in it. Spending caps live in the server (INSUMER_MAX_PAYMENT_USDC for x402 pay-per-call) and in InsumerAPI's credit balance.
  • Only the 18 listed tool names are reviewed. Read-only lookups and any tool added in a later mcp-server-insumer release inherit Guard's usual handling until the entry is updated.
  • The npx launch only. The hosted endpoint at api.insumermodel.com/mcp is reached by URL and is not covered.

Residual effects

Unknown / self-reported

Contributor-supplied residual-effect text is not independently verified. With the coverage active, an approved call still does what it does. An approved insumer_attest or trust call spends the credit or the x402 USDC payment; Guard asks, it does not refund. An approved merchant or domain tool changes that merchant at InsumerAPI, and an approved buy tool registers the payment for good. Approving is the decision. Two things happen with no prompt at all: the read-only lookups run freely, and some of them send a wallet address or a condition to InsumerAPI; and the server's own x402 pay-per-call behavior, including its quote check and cap, runs inside the server, outside Guard's view.

  • An approved call still spends or changes: credits or USDC are taken, merchants are changed, payments are registered. Guard asks; it does not refund or undo.
  • Read-only lookups run with no prompt, and some send a wallet address or a condition to InsumerAPI.
  • x402 pay-per-call quoting, the payee check and the per-call cap run inside the server, outside Guard's view.

Release evidence

Unknown / self-reported

Contributor-supplied release text is not independently verified. The catalog entry is version 1.0.0 and covers the server by tool name, so it does not change when the server releases. The server itself is mcp-server-insumer on npm; the catalog entry was written against its tool list in October 2026 and lists 27 tools. Check the npm page for the current release before relying on a tool count.

Controlled fixture proof (separate)

Unknown / self-reported

No controlled fixture evidence has been supplied.

Exact installation and activation guide

Commands are copied only when the reviewed source supplies an exact argv contract. Other guidance is labeled unknown and is not executed by this page.

Install

Unknown / self-reported

Contributor-supplied install guidance is not independently verified because an exact reviewed argv contract is unavailable. Installing the server is one config entry in your MCP client; enabling the Guard coverage is a separate step a local admin takes in Guard. The steps below are the documented path from the server's README (https://github.com/insumerapi/mcp-server-insumer#readme). Nothing here is run for you.

  1. 1

    Get a key, or skip to step 2 for pay-per-call. One request creates a free key: POST https://api.insumermodel.com/v1/keys/create with a JSON body of {"email": "[email protected]", "appName": "MCP Server", "tier": "free"}. The response carries an insr_live_... key. One free key per email; 10 free verifications plus 100 requests a day.

    Self-reported / unavailable
  2. 2

    Add the server to your MCP client. In Cursor's MCP settings, or in Claude Code's MCP configuration, add a server named insumer with command npx, args ["-y", "mcp-server-insumer"], and env INSUMER_API_KEY set to your key. That runs the latest release, which is where fixes and hardening arrive. If your policy requires pinning, use [email protected] instead; either way, a tool added in a later release inherits Guard's usual handling until this entry is updated, so read the release notes when you move. For x402 pay-per-call, set INSUMER_PAYMENT_KEY to a throwaway Base wallet's private key instead of a key; USDC only, no ETH needed, and optionally INSUMER_MAX_PAYMENT_USDC to lower the $3.00 per-call cap.

    Self-reported / unavailable
  3. 3

    Restart the client and confirm the insumer server lists its 27 tools, each signed answer carrying both an ES256 (ECDSA) signature and a post-quantum ML-DSA-65 signature you can check against https://insumermodel.com/.well-known/jwks.json. A read-only lookup such as fetching those signing keys needs no credit and proves the connection.

    Self-reported / unavailable
  4. 4

    Install HOL Guard if it is not already running (https://hol.org/guard), then continue to Activation to enable this coverage; it stays off until a local admin turns it on.

    Self-reported / unavailable

Activation

Unknown / self-reported

Contributor-supplied activation guidance is not independently verified because an exact reviewed argv contract is unavailable. The catalog declares this coverage external and opt-in. Installing the server does not turn it on, and this page cannot turn it on. A local admin enables mcp.insumer from the Guard workspace controls; review for the 18 listed tools applies only after that local-admin enable, never from a signed-cloud enable. You know it is active when a covered call, such as insumer_attest, stops at Guard's review prompt instead of going straight through, and Guard's activity view records the decision.

  1. 1

    As a local admin, open your Guard workspace controls and find the MCP coverage for mcp.insumer (Guard enforcement ID command.mcp-insumer). Enable it. Guard's own guide for enabling external extensions is at https://github.com/hashgraph-online/hol-guard/blob/main/docs/guard/extensions/README.md.

    Self-reported / unavailable
  2. 2

    Confirm the 18 listed tools now show Review in Guard's policy view for this extension, and that the other entry still reads Inherit policy. No tool should read Allow.

    Self-reported / unavailable
  3. 3

    To stop: disable the extension from the same controls. The server keeps working; its calls simply return to Guard's usual handling. Removing the mcpServers entry from your client removes the server itself.

    Self-reported / unavailable

Safe fixture inputs

Unknown / self-reported

Contributor-supplied safe fixture input guidance is not independently verified because an exact reviewed argv contract is unavailable. The safe trial uses a free key and a wallet you control, so the most it can cost is one of the key's 10 free verifications, and it changes nothing anywhere. It runs one read-only lookup (passes through) and one covered call (should stop for review). Do not use a merchant tool, a buy tool or insumer_setup for the trial; those change state or register payments.

  1. 1

    Ask your agent to fetch InsumerAPI's signing keys through the insumer server. This is a read-only lookup: it costs nothing and is not in the reviewed list, so Guard should let it through without a prompt.

    Self-reported / unavailable
  2. 2

    Ask your agent to run insumer_attest for a wallet address you control with one condition: type token_balance, chainId 8453 (Base), contractAddress 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913 (USDC), operator gte, threshold "1" as a string. Whether the wallet holds a dollar of USDC does not matter; a signed no is as good a result as a signed yes.

    Self-reported / unavailable
  3. 3

    When Guard's review prompt appears for insumer_attest, approve it once. Then open Guard's activity view and find the recorded decision for command.mcp-insumer.

    Self-reported / unavailable

Disable / remove

Unknown / self-reported

Contributor-supplied disable/remove guidance is not independently verified because an exact reviewed argv contract is unavailable. Two separate things can be turned off. Disabling the coverage in Guard stops the review prompts; the server keeps working and its calls go back to Guard's usual handling. Removing the server's entry from your MCP client removes the server. Neither step touches anything at InsumerAPI: your key, its remaining credits, any merchant you created and any payment you registered stay exactly as they were, and any attestation already returned stays valid until its own expiresAt.

  1. 1

    To stop the review prompts: as a local admin, disable mcp.insumer in your Guard workspace controls. The server keeps running; its calls return to Guard's usual handling.

    Self-reported / unavailable
  2. 2

    To remove the server: delete the insumer entry from mcpServers in your MCP client's config and restart the client. npx keeps no install to uninstall.

    Self-reported / unavailable
  3. 3

    What stays behind: your InsumerAPI key and credits, any merchant or configuration created through the tools, and any payment registered. Manage those at https://insumermodel.com/developers/account/ or through the API; Guard never changes them.

    Self-reported / unavailable

FAQ

Frequently asked questions

Reviews the InsumerAPI tools that can spend money or credits or change merchant state: verification and trust calls that use credits or pay per call in USDC, payment registrations, API key and merchant creation, merchant configuration, domain verification and the public listing. Read-only lookups keep Guard's usual handling. Off until you turn it on. The listing declares 0 rules and 1 permission check, with per-tool mapping for 19 tools. No per-command matchers are published for this listing.. Coverage is limited to these reviewed operations and the surrounding Guard policy.

External · opt-in: External coverage. It must be enabled through Guard controls; this page does not activate it. Enabling state is always controlled through Guard policy, never from this directory.

This listing entered the catalog as a community contribution through a public pull request with recorded provenance, with public credit to @douglasborthwick-crypto, and 1 GitHub maintainer is recorded on the listing. Credit is attribution only. A verified publisher profile and the claim flow verify authority separately; neither is an upstream endorsement or a HOL safety certification.

No. Every listing documents source, activation model, maintainer identity, and stated limitations so you can evaluate coverage before enabling it. Review the stated limitations and the exact source tree before relying on any single control.

This listing does not publish per-command matchers. Exact behavior follows your Guard policy combined with this entry's 0 rules and 1 permission check.

External coverage. It must be enabled through Guard controls; this page does not activate it. Enablement is managed through your workspace's Guard policy and controls — never from this directory. Open the install guidance for the setup flow, then adjust the MCP coverage for mcp.insumer in Guard's policy surface.

The matching action is stopped at Guard's pre-action boundary before it executes, and the decision is recorded with evidence your workspace can review. Exact behavior follows your Guard policy combined with this entry's 0 rules and 1 permission check.

No. Guard is local-first: interception, decisions, and evidence stay on your machine unless your workspace explicitly configures cloud features. This page is documentation only — it never executes a command or changes protection state.

Guard runs locally alongside the major AI coding agents and MCP-capable harnesses, so this coverage applies wherever Guard intercepts actions. See the supported harness guides at https://hol.org/guard/harnesses for per-tool approval behavior and limitations.

Guard has a free local tier that includes command interception, evidence, and the policy controls this listing documents. Team plans add shared policy, review queues, and audit surfaces. Current plans: https://hol.org/guard/pricing

Open an issue or pull request against the hol-guard repository, where the canonical catalog lives: https://github.com/hashgraph-online/hol-guard Listings are corrected through the same public review process that adds them.

Yes. Community extensions are merged through public pull requests with recorded provenance, and you can claim the publisher page for a contribution you maintain through the Publisher Studio.

The command mapping table is the audit trail for covered commands and catalog default floors. Pair it with the reviewed rule and permission counts, the per-tool state mapping, and the exact source tree linked from View exact source at the reviewed commit. Guard records enforcement decisions with evidence locally, so what ships matches what you reviewed.