AegisAI

by kadopi · remote · no sign-in to list tools · v0.6.0 · published by GitHub user kadopi

Paid pre-execution risk verification for AI agents over MCP and x402.

Deletes · 1Final · 1Changes · 9

Compare Registry entry

17 tools: 1 can delete or cancel things, 1 do something that can't be undone, 9 change things and 6 only read.

Scanned Oct 21, 2026

Deletes or cancels1Claude Code: deny: see these tools
Can't be undone1Claude Code: ask: see these tools
Changes things9Claude Code: ask: see these tools
Only reads6Claude Code: allow: see these tools

2 checks matched something in the tool definitions. No finding is not a safety guarantee. How the checks work.

claude mcp add --transport http aegis-ai https://aegis-api.kadopi.workers.dev/mcp

Findings

Matches of fixed checks, with the text they matched.

2 warn

Warnevaluate_agent_commerce_policy
The server marks "evaluate_agent_commerce_policy" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
Matched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "evaluate_agent_commerce_policy" says "evaluate": a word for a change.
What this check looks for
Warnverify_risk
The server marks "verify_risk" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
Matched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":true}; The description's first verb is "run": a word for a change.
What this check looks for

Tools

Open a tool for its reasons and full description. Descriptions are the publisher's own words, shown as they are.

17 tools

Deletes or cancels 1Removes, cancels or revokes something.Claude Code: deny

stop_monitor_subscriptionIdempotently disables monitoring for one account-owned subscription without deleting its audit history.No findings
Why it is in this group
- The name "stop_monitor_subscription" says "stop": a word for a change.
- The server marks it destructive (destructiveHint: true).
Description, as the server wrote it
Idempotently disables monitoring for one account-owned subscription without deleting its audit history.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: true
- openWorldHint: false

Can't be undone 1Sends, pays, publishes or merges: once done, it cannot be taken back.Claude Code: ask

evaluate_agent_commerce_policyDeterministically pre-evaluates a public-price sale without sending, paying, settling, or changing stored state.1 finding
Why it is in this group
- The name "evaluate_agent_commerce_policy" says "evaluate": a word for a change.
- The input carries an amount (input.offer.price.amount) and a currency, so it moves money.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
  The server marks "evaluate_agent_commerce_policy" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
  Matched text
  annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "evaluate_agent_commerce_policy" says "evaluate": a word for a change.
  What this check looks for
Description, as the server wrote it
Deterministically pre-evaluates a public-price sale without sending, paying, settling, or changing stored state. Signed Basic and Pro public terms may be allowed; verified-payer velocity is reserved atomically after facilitator verification and before settlement.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false

Changes things 9Creates or changes something that can be changed back.Claude Code: ask

create_monitor_a2a_dns_txt_challengeCreates a one-time DNS TXT ownership challenge without contacting the customer endpoint.No findings
Why it is in this group
- The name "create_monitor_a2a_dns_txt_challenge" says "create": a word for a change.
Description, as the server wrote it
Creates a one-time DNS TXT ownership challenge without contacting the customer endpoint. Verification and delivery remain disabled.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
create_monitor_a2a_endpoint_challengeCreates a one-time 15-minute challenge for a pending endpoint.No findings
Why it is in this group
- The name "create_monitor_a2a_endpoint_challenge" says "create": a word for a change.
Description, as the server wrote it
Creates a one-time 15-minute challenge for a pending endpoint. It does not contact or verify the endpoint and never enables delivery.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
register_monitor_a2a_endpointRegisters an account-scoped endpoint in pending-verification state.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Description, as the server wrote it
Registers an account-scoped endpoint in pending-verification state. Registration never contacts the endpoint and delivery remains disabled.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
register_monitor_subscriptionRegisters a Base Sepolia wallet and validated policy in stopped state.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Description, as the server wrote it
Registers a Base Sepolia wallet and validated policy in stopped state. Payment and operator approval are still required before activation.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
stop_monitor_a2a_endpointIdempotently stops an account-owned endpoint and cancels only its unsent deliveries.No findings
Why it is in this group
- The name "stop_monitor_a2a_endpoint" says "stop": a word for a change.
Description, as the server wrote it
Idempotently stops an account-owned endpoint and cancels only its unsent deliveries.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
update_monitor_alert_statusIdempotently updates only the case status and note for an alert owned by the authenticated monitor account.No findings
Why it is in this group
- The name "update_monitor_alert_status" says "update": a word for a change.
Description, as the server wrote it
Idempotently updates only the case status and note for an alert owned by the authenticated monitor account. It never changes notification state or performs wallet actions.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
update_monitor_subscriptionIdempotently updates label or policy.No findings
Why it is in this group
- The name "update_monitor_subscription" says "update": a word for a change.
Description, as the server wrote it
Idempotently updates label or policy. It never activates monitoring, pays, or changes networks.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
verify_monitor_a2a_dns_txt_challengeChecks only Cloudflare's fixed DNS resolver when DNS verification is enabled.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Description, as the server wrote it
Checks only Cloudflare's fixed DNS resolver when DNS verification is enabled. It never contacts the customer endpoint and never enables delivery.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
verify_riskRuns paid AegisAI risk verification and never runs verification for free.1 finding
Why it is in this group
- The description's first verb is "run": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
  The server marks "verify_risk" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
  Matched text
  annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":true}; The description's first verb is "run": a word for a change.
  What this check looks for
Description, as the server wrote it
Runs paid AegisAI risk verification and never runs verification for free. An x402-aware MCP client can pay and retry this tool automatically; unpaid calls return the canonical payment challenge.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true

Only reads 6Looks things up and changes nothing.Claude Code: allow

get_aegis_capabilitiesReturns machine-readable discovery links and x402 testnet payment information.No findings
Why it is in this group
- The name "get_aegis_capabilities" says "get": a word for reading.
Description, as the server wrote it
Returns machine-readable discovery links and x402 testnet payment information. This tool is free and does not perform risk verification.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
get_monitor_alertReturns one structured alert and its bounded evidence for the authenticated monitor account.No findings
Why it is in this group
- The name "get_monitor_alert" says "get": a word for reading.
Description, as the server wrote it
Returns one structured alert and its bounded evidence for the authenticated monitor account.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
get_monitor_subscriptionReturns one account-owned subscription and its current policy and scan state.No findings
Why it is in this group
- The name "get_monitor_subscription" says "get": a word for reading.
Description, as the server wrote it
Returns one account-owned subscription and its current policy and scan state.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
list_monitor_a2a_endpointsLists the authenticated account's A2A notification endpoints and delivery-disabled state.No findings
Why it is in this group
- The name "list_monitor_a2a_endpoints" says "list": a word for reading.
Description, as the server wrote it
Lists the authenticated account's A2A notification endpoints and delivery-disabled state.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
list_monitor_alertsLists structured alerts belonging only to the authenticated monitor account.No findings
Why it is in this group
- The name "list_monitor_alerts" says "list": a word for reading.
Description, as the server wrote it
Lists structured alerts belonging only to the authenticated monitor account. This tool never sends notifications or performs wallet actions.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
list_monitor_subscriptionsLists account-scoped Base Sepolia monitor subscriptions without changing monitoring or payment state.No findings
Why it is in this group
- The name "list_monitor_subscriptions" says "list": a word for reading.
Description, as the server wrote it
Lists account-scoped Base Sepolia monitor subscriptions without changing monitoring or payment state.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Install

Pinned to version 0.6.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http aegis-ai https://aegis-api.kadopi.workers.dev/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json

{
  "permissions": {
    "allow": [
      "mcp__aegis-ai__get_aegis_capabilities",
      "mcp__aegis-ai__list_monitor_alerts",
      "mcp__aegis-ai__get_monitor_alert",
      "mcp__aegis-ai__list_monitor_subscriptions",
      "mcp__aegis-ai__get_monitor_subscription",
      "mcp__aegis-ai__list_monitor_a2a_endpoints"
    ],
    "ask": [
      "mcp__aegis-ai__evaluate_agent_commerce_policy",
      "mcp__aegis-ai__update_monitor_alert_status",
      "mcp__aegis-ai__register_monitor_subscription",
      "mcp__aegis-ai__update_monitor_subscription",
      "mcp__aegis-ai__register_monitor_a2a_endpoint",
      "mcp__aegis-ai__create_monitor_a2a_endpoint_challenge",
      "mcp__aegis-ai__create_monitor_a2a_dns_txt_challenge",
      "mcp__aegis-ai__verify_monitor_a2a_dns_txt_challenge",
      "mcp__aegis-ai__stop_monitor_a2a_endpoint",
      "mcp__aegis-ai__verify_risk"
    ],
    "deny": [
      "mcp__aegis-ai__stop_monitor_subscription"
    ]
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. It assumes the server is installed as aegis-ai. Review it before you use it.

Keep track

Get told when this server's tools change: a new tool, a new description, a new class.

https://anlyon.dev/servers/io.github.kadopi/aegis-ai/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Anlyon alternative

11 of these tools change something. Run those as governed actions.

Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.

Anlyon does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.

Import an OpenAPI spec →

Compare with other servers for the same product.

Source: MCP Registry. Synced Oct 21, 2026, 18:37 UTC.

For agents: Markdown · JSON · report this listing