agentlab-services

Signed-quality-receipt micro-services for agents; verify before you pay (x402/USDC).

使うべきか

品質と安全性

B
説明の品質
82%
スキーマの完全性
30%
命名の品質
83%
ポイズニングのリスク
100%
権限の一致
100%
プロトコルへの準拠
100%

検出事項(5)

  • LOWTool 'market_brief' description lacks action verbmarket_brief 内
  • LOWTool 'screen_message' description lacks action verbscreen_message 内
  • LOWTool 'clean_csv' description lacks action verbclean_csv 内
  • LOWTool 'honesty_audit' description lacks action verbhonesty_audit 内
  • LOWTool 'trust_lease' description lacks action verbtrust_lease 内

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

~698トークン数(ツール定義)
~183 B一般的なレスポンスサイズ
注意への影響は中程度(128k コンテキストの 0.55%)

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": {
    "agentlab-services": {
      "url": "https://sigmaintelligence.online/mcp"
    }
  }
}

リモートエンドポイント

https://sigmaintelligence.online/mcpstreamable-http

できること

ツール一覧

ツール(14)

🟢 読み取り専用🟡 書き込み🔴 削除⚪ 不明
⚪market_brief

Evidence-backed demand/pain/opportunity brief for a topic (HN + StackExchange + GitHub signals).

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪screen_message

Phishing / social-engineering risk score for a message.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
🟢verify_text

Deterministic quality check of a deliverable against a spec (length, coverage, honesty, structure).

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪clean_csv

Clean, dedupe and normalize a CSV; returns a tidy CSV.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪rank_opportunities

Rank candidate money/leads by euros-per-hour and speed-to-cash toward a target.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪audit_opportunity

Conservative fit, fee, scam, eligibility and pay audit for a work opportunity.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪productize_service

Turn a raw capability into a fixed-scope, priced, machine-readable commercial offer.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪notarize

Proof-of-existence: signed, timestamped receipt binding any content to a hash (IP/plagiarism/audit).

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪honesty_audit

Honesty linter for AI output: flags unsupported certifications, fabricated authority and uncited statistics.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪endpoint_safety

Verify-before-you-pay: pre-flight scam/liveness/https screen of an x402 or MCP endpoint.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
🟢attest_quality

QA notary: independently quality-check ANOTHER agent's deliverable and issue a signed pass/fail.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪trust_lease

24-hour renewable trust passport for an agent/API: identity surfaces, schema availability and drift, bound to a signed receipt.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪revenue_leak

Revenue-leak diagnostic for an agent-economy seller: which discovery surfaces are missing and whether a buying agent can settle machine-to-machine at all.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}
⚪revenue_watch

30-day recurring revenue-leak monitor: reports newly appeared and resolved leaks against the last settled baseline, bound to a signed receipt.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": true
}

コミュニティ

このサーバーを評価する

エビデンス

最近の観測

検証済みバージョンは記録されていませんツール 14 件