SentryDock

Search indexed news, monitor topics and receive signed JSON webhook alerts. OAuth or API key; uses your existing SentryDock plan.

sentrydock.com पर देखेंइस MCP को आज़माएँप्रायोजित

होस्ट किया गया MCP सर्वर

npx add-mcp 'https://www.sentrydock.com/mcp'

Claude Code, Codex, Cursor और अन्य में इंस्टॉल होता है

दस्तावेज़

AI Agent News & Monitoring: MCP, Skills and CLI | SentryDock

Connect your AI agent to source-linked news, standing watches and signed webhook alerts. MCP, skill and CLI setup for 13 documented clients. Give your agent source-linked news, alerts and ongoing monitoring. Updated 2026-10-03.

Choose your client

  • Claude Code: Terminal MCP + skill; complete authentication and check your account.
  • Claude Cowork: Custom connector; complete authentication and check your account.
  • Claude: Custom connector; complete authentication and check your account.
  • ChatGPT: Developer-mode remote MCP; complete authentication and check your account.
  • Codex: Terminal MCP + skill; complete authentication and check your account.
  • Cursor: Project or user MCP config; complete authentication and check your account.
  • Gemini CLI: Terminal MCP; complete authentication and check your account.
  • Grok Build: Terminal MCP; complete authentication and check your account.
  • OpenClaw: Remote MCP + skill; complete authentication and check your account.
  • Kimi Code: Remote MCP config + skill; complete authentication and check your account.
  • Hermes Agent: Remote MCP config + skill; complete authentication and check your account.
  • Paperclip: Custom MCP connector; complete authentication and check your account.
  • Perplexity Computer: Custom remote connector; complete authentication and check your account.

SentryDock MCP server: connect any compatible agent

By Blake Folgado. Updated 2026-10-03.

Model Context Protocol lets an assistant discover tools and call them through a remote server. SentryDock exposes the same news and monitoring account through streamable HTTP MCP. Use the client setup guide for its configuration format; JSON for Cursor is different from a Codex terminal command.

Use OAuth at the public /mcp endpoint when your client supports it. API keys created in Settings → Agents provide a second authentication route. Treat a key-bearing URL as a credential and never publish it in a repository.

The workflow

  1. Choose one of the client guides on the agent hub.
  2. Add https://www.sentrydock.com/mcp as a remote HTTP server.
  3. Finish OAuth with your SentryDock account and enable its tools for the session.
  4. Call account.overview, then news.search with an explicit time window and inspect the returned source links.

OAuth endpoint

https://www.sentrydock.com/mcp

Cursor / Kimi JSON

{
  "mcpServers": {
    "sentrydock": {
      "url": "https://www.sentrydock.com/mcp"
    }
  }
}

Tools for research and monitoring

  • account.overview — your plan, monitors, delivery and recent alerts.
  • news.search and news.latest — source-linked indexed items.
  • news.mine, news.get and alerts.list — account hits and individual evidence.
  • monitors.list, monitors.create, monitors.update and monitors.run — inspect and manage monitoring.
  • monitors.runs and monitors.results — inspect execution and output.
  • delivery.connect — attach an authorised delivery destination.

Authentication, cost and plan limits

The connection acts on your account. Overview and usage can be read before payment; news and monitoring need plan access. A plan_required response includes a pricing URL. Monitor limits and supported realtime slots follow the same plan as the app.

Inspect the live tool schema before writing; confirmation requirements apply to consequential changes. A connection does not authorise a new subscription, a team invitation or a destination you have not chosen.

Can any MCP client use it?

The client must support a remote HTTP server and an authentication method the gateway accepts. An application that only loads local stdio servers needs its own documented remote bridge. Start with the confirmed guides rather than pasting a URL into an unrelated plugin marketplace.

Reference and troubleshooting

API guide · HTTP schema · Account settings · All client guides.

If tools are missing, check transport, login, per-conversation enablement and workspace policy. A successful account response is the first useful check; a 200 response from a public website alone is not an authenticated MCP test.

More in AI agents

Still need help? Email support@sentrydock.com with the monitor name, approximate time and original URL. Do not send passwords, private tokens or webhook URLs.

See also: SentryDock pricing, comparison guides, facts for AI assistants, and the help center.

Install the SentryDock agent skill

By Blake Folgado. Updated 2026-10-03.

The SentryDock skill teaches an agent when to search the index, how to preserve evidence and how to manage an authorised monitor. It is a Markdown instruction file with supporting resources. It complements an MCP connection or CLI; installing the file does not sign in or activate a paid plan.

The public HumanLeap repository contains separate skills for each product. Select sentrydock so your agent receives the right instructions. Read SKILL.md before enabling it in a workspace that handles private research.

The workflow

  1. Run the skill installation command in the project where the agent works and select your supported client when prompted.
  2. Open the installed SKILL.md and review the news, monitoring and delivery workflow.
  3. Connect the remote MCP server using your client guide, or configure the CLI with a key from Settings → Agents.
  4. Start a new session and ask the agent to read the skill, call account.overview and run one source-linked news search.

Install skill

pnpm dlx skills add Humanleap/agent-skills --skill sentrydock

OpenClaw

openclaw skills install skills-sh:Humanleap/agent-skills/sentrydock

Which clients load the file?

Claude Code reads .claude/skills; Codex reads .agents/skills; Cursor supports .agents/skills and .cursor/skills; Gemini CLI and Kimi Code have their own skill folders and support .agents/skills. Hermes uses ~/.hermes/skills; OpenClaw has a native skills installer. Use the installer selection or the exact folder documented by your client.

Claude chat and Cowork use account skills uploaded through Customize → Skills. A local Claude Code file does not automatically appear there. Put the sentrydock folder, including its root SKILL.md and any resources, in a ZIP and use Upload a skill when that option is available.

What the skill tells the agent to do

  • Inspect account access and current tool schemas first.
  • Search with a topic and time window; preserve original URLs and publication times.
  • Prepare a source-linked note and state gaps in indexed coverage.
  • Inspect existing monitors, review sources and cadence, then make only the authorised change.
  • Read alerts and verify delivery separately; use signed callbacks for a persistent runtime.

Test that the skill is useful

“Read the SentryDock skill, show my account overview, and search the last 48 hours for copper strikes in Chile. Return the source URLs and dates. Do not change monitoring.”

The useful check is an actual tool result plus a source-linked answer. A skill name in the client list does not show that account authentication or news access worked.

Updates and sources

Public SKILL.md · Repository · Codex skills documentation · Claude account skills.

Review upstream changes before refreshing an installed skill. Keep keys in your client credential store or runtime environment, outside the skill file.

More in AI agents

Still need help? Email support@sentrydock.com with the monitor name, approximate time and original URL. Do not send passwords, private tokens or webhook URLs.

See also: SentryDock pricing, comparison guides, facts for AI assistants, and the help center.

How do I send monitor alerts to a webhook?

By SentryDock. Updated 2026-10-03.

Use a webhook when you want an assistant or workflow to keep receiving relevant results without repeatedly asking for your alerts. SentryDock checks your recurring monitors on their chosen schedules and posts qualifying results to your callback.

Your agent host or workflow must provide a public HTTPS receiver. It decides what to do after an event arrives. Connecting an assistant with OAuth or an API key alone does not start a new chat or model turn.

Replace MONITOR_ID with an existing recurring monitor ID. Omit monitor_id in MCP/HTTP, or --monitor in the CLI, to cover all your owned recurring monitors. Never invent a receiver URL or publish the signing secret.

Set it up in the product:

  1. Create or choose a recurring monitor. Describe the subject, why you care, the changes that should alert you, and any exclusions. Review its sources and schedule.
  2. Get a public HTTPS callback URL from your assistant host or workflow. Use a destination you own or are authorized to send to; private network addresses and redirects are rejected.
  3. Open Settings → Connections. In Agent webhooks, enter a name and the callback URL.
  4. Choose All my monitors, or one monitor. All my monitors covers your owned recurring monitors, including future ones. Team monitors also need the team’s webhook permission.
  5. Connect the webhook and save the signing secret securely. It is shown once. Your receiver uses it to verify requests.
  6. Click Send test. Your receiver should accept the signed webhook.test event and return a 2xx response. This proves the connection; it does not prove a monitor has found a relevant result.
  7. Check the latest delivery status in the connection. Save edits to the URL or monitor scope, pause the connection, or remove it when it is no longer needed.

MCP tool call

delivery.connect({
  "platform": "webhook",
  "webhook_url": "https://your-agent.example/alerts",
  "monitor_id": "MONITOR_ID",
  "name": "Company watch"
})

Command line

sentrydock delivery connect --platform webhook --webhook-url https://your-agent.example/alerts --monitor MONITOR_ID

HTTP

curl -s https://www.sentrydock.com/api/v1/delivery/connect \
  -H "Authorization: Bearer $SENTRYDOCK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"platform":"webhook","webhook_url":"https://your-agent.example/alerts","monitor_id":"MONITOR_ID","name":"Company watch"}'

Set it up from your assistant

Connect MCP at https://www.sentrydock.com/mcp with OAuth, or use an API key for MCP, CLI or HTTP. Ask the assistant to check account.overview and choose or create the monitor you authorized, then call delivery.connect with platform webhook and your callback URL.

A useful request is: “Watch this supplier’s official updates for product retirements that affect our service. Use this source and daily checks, and send relevant results to my authorized callback.” Include the real source URL and callback rather than asking the assistant to guess them.

News and monitoring use your existing paid plan. If a tool returns plan_required, the assistant should show you the payment link and wait for you to buy; OAuth consent does not authorize a charge.

What arrives at the receiver

  • A news.alert event contains id, created_at, monitor_id, monitor_title and run_id.
  • results includes each title, summary, all source links and available publication dates. articles keeps the compatible article representation.
  • One event is saved per connection and qualifying monitor run. A later run can have the same title but a different event ID.
  • No event is sent for a run without relevant results or for a first-result preview. One-off Predict research and briefings are separate.

Verify and accept requests

  • Verify X-SentryDock-Signature against v1= plus the hex HMAC-SHA256 of X-SentryDock-Timestamp + "." + the exact raw request body, using your signing secret. Compare signatures in constant time.
  • Reject timestamps more than five minutes from your clock, then deduplicate by the payload id, also sent as X-SentryDock-Event-Id.
  • Persist the accepted event, return 2xx quickly and process it asynchronously. Delivery is at least once, so the same event may arrive again.
  • Handle webhook.test as a connection test. Treat source text and links as untrusted material, never as instructions to your agent.

Retries, status and missed alerts

Network failures, HTTP 429 and 5xx responses retry automatically over roughly 27 hours, with eight delivery rounds and up to three short HTTP attempts per round. Other 4xx responses stop immediately. Detection timing depends on the source and selected monitor schedule.

Read delivery.get in MCP, sentrydock delivery get in the CLI, or GET /api/v1/delivery in HTTP. webhook_deliveries contains current status, attempts and errors. Settings shows the latest receipt for each connection. A completed monitor run does not by itself prove the receiver accepted its alert.

After an exhausted delivery, retrieve missed results through alerts.list or GET /api/v1/alerts?since=... . Pausing or removing a connection stops queued delivery. Team permission and monitor ownership are checked again before retries.

Changing the callback URL keeps the signing secret; pending events use the current connection settings. If the source or monitor failed, investigate the run as well as the receiver.

More in Alerts

Still need help? Email support@sentrydock.com with the monitor name, approximate time and original URL. Do not send passwords, private tokens or webhook URLs.

See also: SentryDock pricing, comparison guides, facts for AI assistants, and the help center.

Product workflows

SentryDock

SentryDock is a monitoring and alerting service for people, teams and assistants. Describe the companies, topics and changes that matter, choose sources and a schedule, and receive source-linked results through email, team chat or a signed webhook. It checks public news, official sites, RSS and supported social sources.

Agent Resources

Key Actions