AIDNode | AI Agent Support Network
Five AI-agent MCP tools: Repair, Verify, Witness, Monitor, Recovery. Free beta; no billing.
使うべきか
品質と安全性
ツール定義とプロトコルへの準拠に関する自動分析に基づいています。
コンテキストコスト
これは、サーバーのツールがモデルのコンテキストに読み込まれるたびに消費されるおおよそのトークン数です。数が多いほど、ほかのタスクに使える注意が減ります。
インストール
ワンクリックインストール
これを `claude_desktop_config.json` ファイルに追加してください:
{
"mcpServers": {
"aidnode": {
"url": "https://botstop-dev.botstop-dev-lab-2026.workers.dev/mcp"
}
}
}リモートエンドポイント
https://botstop-dev.botstop-dev-lab-2026.workers.dev/mcpstreamable-httpできること
ツール一覧
ツール(6)
⚪repair(operation_id, status, network_error, method, attempt, ...)
AIDNode Repair v2. Use after an HTTP or network call fails, before retrying: classifies the failure (status code or transport error), decides whether an automatic retry is safe given method idempotency and Retry-After, and returns a bounded backoff plan. Advisory only; AIDNode never sends the request. For multi-step workflow failures use recovery instead.
入力スキーマ
{
"type": "object",
"properties": {
"operation_id": {
"type": "string",
"description": "Unique call id 8-100 characters; counted once by beta ledger"
},
"status": {
"type": "integer",
"minimum": 100,
"maximum": 599
},
"network_error": {
"type": "string",
"enum": [
"ECONNRESET",
"ETIMEDOUT",
"ECONNREFUSED",
"EAI_AGAIN",
"DNS_ERROR",
"NETWORK_ERROR",
"TLS_ERROR"
]
},
"method": {
"type": "string",
"enum": [
"GET",
"HEAD",
"OPTIONS",
"PUT",
"DELETE",
"POST",
"PATCH"
]
},
"attempt": {
"type": "integer"
},
"max_attempts": {
"type": "integer"
},
"idempotency_key_present": {
"type": "boolean"
},
"request_may_have_committed": {
"type": "boolean"
},
"response_headers": {
"type": "object"
},
"retry_after_seconds": {
"type": "number"
}
},
"required": [
"operation_id"
],
"additionalProperties": false
}🟡verify(operation_id, mode, previous, current, url, ...)
AIDNode Verify v2. Use to check whether content changed: compares two supplied texts (hash, first changed index) or two JSON values (canonical change paths), or fetches a small artifact from a fixed allowlist of public HTTPS hosts and returns its digest. Does not judge factual truth. For endpoint uptime/status use monitor instead.
入力スキーマ
{
"type": "object",
"properties": {
"operation_id": {
"type": "string",
"description": "Unique call id 8-100 characters; counted once by beta ledger"
},
"mode": {
"type": "string",
"enum": [
"compare_text",
"compare_json",
"public_http"
]
},
"previous": {},
"current": {},
"url": {
"type": "string"
},
"expected_sha256": {
"type": "string"
}
},
"required": [
"operation_id"
],
"additionalProperties": false
}⚪witness(operation_id, document, required_fields, expected_values, expected_types, ...)
AIDNode Witness v2. Use to produce a tamper-evident record that a supplied JSON document met explicit checks (required fields, expected values/types, field_rules): returns pass/fail per assertion and an HMAC-signed receipt verifiable at /receipts/verify. Does not attest that a real-world action happened.
入力スキーマ
{
"type": "object",
"properties": {
"operation_id": {
"type": "string",
"description": "Unique call id 8-100 characters; counted once by beta ledger"
},
"document": {
"type": "object"
},
"required_fields": {
"type": "array",
"items": {
"type": "string"
}
},
"expected_values": {
"type": "object"
},
"expected_types": {
"type": "object"
},
"field_rules": {
"type": "object",
"additionalProperties": {
"type": "object",
"properties": {
"non_empty": {
"type": "boolean"
},
"min_length": {
"type": "integer",
"minimum": 0,
"maximum": 100000
},
"max_length": {
"type": "integer",
"minimum": 0,
"maximum": 100000
},
"format": {
"type": "string",
"enum": [
"uuid",
"sha256",
"date-time"
]
},
"minimum": {
"type": "number"
},
"maximum": {
"type": "number"
},
"enum": {
"type": "array",
"minItems": 1,
"maxItems": 100
}
},
"additionalProperties": false
}
},
"expected_sha256": {
"type": "string"
},
"external_evidence": {
"type": "object",
"properties": {
"url": {
"type": "string"
},
"expected_sha256": {
"type": "string"
}
},
"required": [
"url",
"expected_sha256"
]
}
},
"required": [
"operation_id",
"document",
"required_fields"
],
"additionalProperties": false
}🟢monitor(operation_id, url, method, expected_status, expected_sha256, ...)
AIDNode Monitor v2. Use to check right now whether an allowlisted public HTTPS endpoint is up: one HEAD/GET probe returning status, latency, health classification and optional content digest match. On-demand only, no schedules. To compare content you already have use verify.
入力スキーマ
{
"type": "object",
"properties": {
"operation_id": {
"type": "string",
"description": "Unique call id 8-100 characters; counted once by beta ledger"
},
"url": {
"type": "string"
},
"method": {
"type": "string",
"enum": [
"HEAD",
"GET"
]
},
"expected_status": {
"type": "array",
"items": {
"type": "integer"
},
"minItems": 2,
"maxItems": 2
},
"expected_sha256": {
"type": "string"
},
"max_duration_ms": {
"type": "integer",
"minimum": 1,
"maximum": 30000
}
},
"required": [
"operation_id",
"url"
],
"additionalProperties": false
}⚪recovery(operation_id, steps)
AIDNode Recovery v2. Use when a multi-step agent workflow failed partway: given steps with status, dependencies, idempotency and side effects, returns which steps to resume in order, which to preserve, which are blocked because their outcome is uncertain, and compensation candidates. Never executes steps. For a single failed HTTP call use repair.
入力スキーマ
{
"type": "object",
"properties": {
"operation_id": {
"type": "string",
"description": "Unique call id 8-100 characters; counted once by beta ledger"
},
"steps": {
"type": "array",
"items": {
"type": "object",
"properties": {
"id": {
"type": "string"
},
"status": {
"type": "string",
"enum": [
"pending",
"running",
"succeeded",
"failed",
"skipped"
]
},
"depends_on": {
"type": "array",
"items": {
"type": "string"
}
},
"idempotent": {
"type": "boolean"
},
"side_effects": {
"type": "boolean"
},
"has_verified_outcome": {
"type": "boolean"
},
"has_compensation": {
"type": "boolean"
}
},
"required": [
"id",
"status"
],
"additionalProperties": false
}
}
},
"required": [
"operation_id",
"steps"
],
"additionalProperties": false
}🟡submit_beta_feedback(checkpoint, service, helpful, would_reuse, would_pay_half_cent, ...)
Optional and free. Report whether AIDNode helped and whether your operator might consider the proposed $0.005/use price. A bot answer is NOT purchase authorization. Feedback is stored privately and is not a testimonial.
入力スキーマ
{
"type": "object",
"properties": {
"checkpoint": {
"type": "integer",
"enum": [
10,
50,
100,
250,
500
]
},
"service": {
"type": "string",
"enum": [
"repair",
"verify",
"witness",
"overall"
]
},
"helpful": {
"type": "string",
"enum": [
"yes",
"no",
"unsure"
]
},
"would_reuse": {
"type": "string",
"enum": [
"yes",
"no",
"unsure"
]
},
"would_pay_half_cent": {
"type": "string",
"enum": [
"yes",
"no",
"unsure"
]
},
"outcome": {
"type": "string",
"enum": [
"resolved",
"partially_resolved",
"not_resolved",
"not_applicable"
]
},
"comment": {
"type": "string",
"maxLength": 500
}
},
"required": [
"checkpoint",
"service",
"helpful",
"would_reuse",
"would_pay_half_cent",
"outcome"
],
"additionalProperties": false
}コミュニティ
エビデンス