Agent Spend Guard — owner-set purchasing guardrails
com.a2awire/agent-spend-guard
All versions
| Version | Registry status |
|---|---|
| 0.1.1latest | Active |
| 0.1.0 | Active |
Version details
- Version
- 0.1.1
- Source
- Source: Official MCP Registry
- Published
- Registry last updated
- First observed by Squatch
- Content fingerprint
- 77f78e09182dc862bef7cdb43f389fa2b9cdf715d7a2d29342a4413ee4cbe119
Transports in this version
- remotestreamable-httphttps://a2awire.com/mcp/connectors/spend-guard/httpNo authentication
Registry metadata (server.json, sanitized)
{
"_meta": {
"io.modelcontextprotocol.registry/official": {
"status": "active",
"isLatest": true,
"updatedAt": "2026-11-08T16:09:30.349601Z",
"publishedAt": "2026-11-08T16:09:30.349601Z",
"statusChangedAt": "2026-11-08T16:09:30.349601Z"
}
},
"server": {
"name": "com.a2awire/agent-spend-guard",
"_meta": {
"io.modelcontextprotocol.registry/publisher-provided": {
"price": "Free. Humans create an API key at a2awire.com/api-key; agents are born — one EIP-712-signed POST to api/v1/birth mints identity + key (signing recipe in llms.txt). Spend Guard never moves money.",
"product": "Spending guardrails for autonomous AI agents. Owners set rules (per-purchase threshold, monthly cap, denylist); the agent calls check BEFORE checkout and gets approve, deny, or ask_user with a plain-English reason.",
"how_it_works": "spend_guard_set_rules -> spend_guard_check_purchase (before checkout) -> spend_guard_record_purchase (after); spend_guard_report for the ledger",
"sample_queries": [
"Merchant acme-store, $8.99 — approved?",
"Set a $25 per-purchase threshold and $200 monthly cap",
"What has this agent spent this month?"
],
"getting_started": [
"connect — no key needed; initialize returns connector instructions",
"get a key — humans: a2awire.com/api-key; agents: be born (one EIP-712-signed POST /api/v1/birth) and use the kit api_key here — the old bootstrap mint is closed (410)",
"spend_guard_set_rules — owner sets threshold, monthly cap, denylist",
"spend_guard_check_purchase — call before checkout; returns approve / deny / ask_user + reason",
"spend_guard_record_purchase — call after checkout; tx_ref makes check+record idempotent"
],
"search_keywords": "[redacted]"
}
},
"title": "Agent Spend Guard — owner-set purchasing guardrails",
"$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
"remotes": [
{
"url": "https://a2awire.com/mcp/connectors/spend-guard/http",
"type": "streamable-http"
}
],
"version": "0.1.1",
"repository": {
"url": "https://github.com/ee324/a2awire",
"source": "github"
},
"websiteUrl": "https://a2awire.com/spend-guard",
"description": "Free spend guardrails for AI agents: approve/deny/ask_user, caps, dupes."
}
}