Mmcp.market

Figuard MCP server

by figuard·io.github.figuard/figuard-mcp·v1.2.0
No reviews yet

Pre-flight spend authorization for AI agents. Set a budget, enforce limits, audit every decision.

B82/100grade B
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
B82/100

full report

Adoption
Growing

2 stars56 downloads/wk

Reviews

Write one

Nobody has reviewed Figuard yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Figuard tools (16, 5 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • figuard_authorize

    Pre-flight authorization — ask FiGuard if a spend is permitted before taking any action.

  • figuard_cancel_batchwrite action

    Cancel one or more budgets. Use this to cancel a single budget (pass one ID) or up to 100 at once.

  • figuard_confirm

    Confirm a previously authorized spend after the action succeeds.

  • figuard_create_budgetwrite action

    Create a new FiGuard budget for an agent session. Returns a session_token that must be passed to figuard_authorize.

  • figuard_create_delegation_tokenwrite action

    Create a scoped delegation token for a fleet budget. Each sub-agent (e.g. per-customer refund agent)

  • figuard_extend_budget

    Extend a budget's expiry window. Use this to keep a long-running agent alive past its original expiry.

  • figuard_fail

    Mark an authorized spend as failed when the action did not succeed (e.g. payment processor declined, API error).

  • figuard_fund_budget

    Adjust a budget's totalLimit in-place without creating a new session.

  • figuard_get_budget

    Get the current state of a budget — total limit, amount spent, amount reserved, amount available, status, and per-category allocations.

  • figuard_get_delegation_token

    Get the current state of a delegation token — label, status, per-category cap usage.

  • figuard_get_ledger

    Query the authorization ledger for a budget — returns a paginated list of spend events (authorized, confirmed, denied, voided, failed).

  • figuard_get_spend_tree

    Get the hierarchical spend tree for a budget — every event and its causal children (parent → child chains).

  • figuard_resume_budget

    Resume a budget that was paused by anomaly detection or manually.

  • figuard_revoke_delegation_token

    Revoke a delegation token immediately. Any subsequent figuard_authorize call using this token

  • figuard_update_budgetwrite action

    Update an existing budget's settings: total limit, velocity caps, expiry, trust mode, or status.

  • figuard_voidwrite action

    Cancel an authorized reservation that was never used.

Public scan report

scanner v0.1.9 · 2026-09-26 · same rubric, same numbers if you re-run it

no findings
  • Code scan7 source files scanned25/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitystatic API keys via environment variables6/15
  • Maintenancelast push 16 days ago15/15
  • Maintainer identityregistry namespace matches repository owner7/10
Overall 82/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

What the publisher says

From the Figuard repository's README, as published. We do not edit it. Read it on GitHub

&color=0A5C38)

A travel-booking agent hit a Stripe timeout and retried twice. The customer's card was charged three times for the same flight before anyone noticed — 40 minutes later.

No alert fired. No limit existed. The agent had a valid API key and no concept of "I already did this."

FiGuard gives agents bounded resources — money, tokens, API calls, GPU hours, any unit you define — and they ask permission before consuming them. You set the ceiling, the retry rules, and the idempotency policy once. Every attempt, authorized or denied, lands in an append-only audit log.

That exact failure is in the stress harness: a retried charge produces 0 double-charges, and 100 agents racing one budget produce 0 overspends — verified against the ledger, reproducibly (make bench).

Your framework decides what to do next. FiGuard decides whether the resource-consuming action is allowed.

Your agent code  (LangChain · LangGraph · CrewAI · any runtime)
  orchestrates — decides what to do next
          ↓  agent wants to spend / call / execute
  figuard.authorize()
  checks: limit · category · velocity · dedup
          ↓  AUTHORIZED — action proceeds
  Stripe · OpenAI · any API or service
  executes — real money or resource consumed
          ↓  action completes
  figuard.confirm()
  settles reservation — ledger updated

LangChain / LangGraph — FiGuard authorizes each tool call before it executes. A budget-exhausted agent stops cleanly instead of running up cost — even across parallel nodes in a LangGraph.

CrewAI — Each crew member gets a delegation token with its own cap. A runaway specialist is stopped at its limit without affecting the rest of the crew.

OpenAI Agents SDK / MCP — Wrap tools with @guardedfunctiontool or add the FiGuard MCP server — every tool call is pre-flight authorized before it reaches the API.

Not using a framework? — The raw SDK works anywhere — a Python script, a background job, a serverless function. If it calls an API that costs money or consumes a bounded resource, FiGuard fits.

Try it now — no setup, no signup: → pip install figuard — runs locally on your machine, nothing hosted (Quickstart) → Run in Colab — or try it in the browser → Live dashboard

FiGuard is the authorization and ledger layer — not a payment processor, not a policy DSL, not an adversarial-agent firewall. Full scope →

Quickstart

Tested with:

pip install figuard
from figuard import FiGuardClient

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Install directly

Runs npx -y figuard-mcp on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add figuard-mcp -- npx -y figuard-mcp
Add to Cursor

Figuard: common questions

Is Figuard MCP server safe?
Mostly: it is graded B (82/100). Read the Figuard safety report
How do I install Figuard?
It runs on your machine. Copy the Claude Code, Claude Desktop or Cursor config from the install section.
Does Figuard need an API key?
Yes. The registry entry asks for FIGUARD_API_KEY.
Is Figuard maintained?
The last commit was 16 days ago (2026-09-11). The latest release is v1.2.0.
What can I use instead of Figuard?
Servers from other publishers that do the same job: Graneth MCP server, Phoenix Zero — L2 Health Oracle MCP server and Gitworthy MCP server. Compare all Figuard alternatives.

Alternatives to Figuard

Same job from other publishers: the closest match first, then the best rated.

All Figuard alternatives →
  • Graneth
    Pre-flight check for AI coding agents: hallucinated packages + secrets, 6 ecosystems, no account.
    A
  • Phoenix Zero — L2 Health Oracle
    Pre-flight L2 health check using kernel-level eBPF telemetry. PASS/DEGRADED/FAIL verdict.
    A
  • Gitworthy
    is it worth your commit? Pre-flight triage for OSS contribution targets.
    B
  • Variflight
    VariFlight's official MCP server provides tools to query flight, weather, comfort, and fare data.
    A
  • Agentic Ledger
    Local-first flight recorder for AI agents: sessions, costs, loop runs, and stuck-loop flags.
    A
Figuard vs GranethFiguard vs Phoenix Zero — L2 Health OracleFiguard vs Gitworthy

More from figuard →