Does the agent search the whole web live?

News search ranks items already indexed by SentryDock. An empty result does not establish that an event did not happen. Check the time window and source coverage, or set up a monitor for the sources you need.

Does monitoring continue after the chat ends?

Yes. A saved monitor runs in your SentryDock account on its configured schedule. The assistant reads alerts when it asks; a persistent agent needs a reachable signed webhook receiver for push delivery.

How quickly do alerts arrive?

The saved cadence and source access determine detection. New monitors default to hourly. Realtime availability and slots depend on the plan; a realtime option is not a guarantee of immediate detection from every source.

The connection uses the same account and plan as the app. Overview and usage can be read before payment; indexed news and monitoring require paid access. Check the pricing page and account limits before running a larger coverage book.

Does the skill grant account access?

No. SKILL.md teaches the workflow. OAuth or a securely configured API key grants access. A local skill does not automatically become a cloud connector or an approved marketplace listing.

Can I control changes?

The agent acts on your account. Give it the exact monitor, source and destination scope, and review changes before authorising them. Remove the connector or revoke its API key to stop access; pause monitors separately to stop scheduled monitoring.

SentryDock for AI agents · Updated 3 October 2026

AI agent news and monitoring with SentryDock

Give your agent source-linked news, your alerts and a coverage book that keeps running after the chat ends. Give your assistant a standing watch: one connection for research, market notes and signed webhook alerts.

Connect your account Install the skill Set up webhook alerts

MCP OAuth, API key or CLI · Same account and plan as the app

Which agents can use SentryDock?

These 13 clients document a remote MCP setup method. Each guide includes the client configuration, authentication, example tasks, ongoing monitoring and troubleshooting. Complete sign-in and check an actual account response to verify your own connection.

Agent clients, connection method and authentication
Agent	Connection	Authentication
Claude CodeAnthropic	Terminal MCP + skill	OAuth sign-in
Claude CoworkAnthropic	Custom connector	OAuth sign-in
ClaudeAnthropic	Custom connector	OAuth sign-in
ChatGPTOpenAI	Developer-mode remote MCP	OAuth sign-in
CodexOpenAI	Terminal MCP + skill	OAuth sign-in
CursorCursor	Project or user MCP config	OAuth sign-in
Gemini CLIGoogle	Terminal MCP	OAuth sign-in
Grok BuildxAI	Terminal MCP	OAuth sign-in
OpenClawOpen source	Remote MCP + skill	OAuth sign-in
Kimi CodeMoonshot AI	Remote MCP config + skill	OAuth sign-in
Hermes AgentNous Research	Remote MCP config + skill	OAuth sign-in
PaperclipOpen source	Custom MCP connector	OAuth sign-in
Perplexity ComputerPerplexity	Custom remote connector	OAuth sign-in

Seven more clients: current setup status

These clients have different connector or runtime requirements. Use the official method available in your account; a product plugin or universal OAuth setup is not confirmed here.

OpenAI dots
Dots uses plugins from the connected ChatGPT account. Use the ChatGPT or Codex guide for its supported tools; a separate product plugin for dots is not confirmed.
Grok Bot
Grok Bot uses its own plugin marketplace and saved workflows. Use the Grok Build guide for terminal MCP; Bot plugin availability needs to be checked in your account.
DeepSeek Harness
HTTP MCP and local skills are documented. Browser OAuth is not documented here; operators need a supported credential flow or a configured CLI runtime.
NanoClaw
Setup is managed by the operator of the group and container. Follow the runtime documentation for skill mounts, MCP plugins and credentials; there is no one-command connection guide here.
Meta Muse
Muse supports custom API connectors. Use the service API documentation and supported account credentials; generic remote MCP and local skill import are not documented here.
Manus
Manus supports custom API or MCP connectors and single-skill imports. A skill package needs SKILL.md at its root. Check the authentication options available in your account.
Cue
Cue has separate connector and skill settings. Skills imported into Manus do not transfer to Cue; check the options in your Cue account before connecting.

How do I connect my agent?

Choose MCP or the CLI. For a compatible client, add the remote MCP server and finish OAuth. For a shell runtime, install the CLI and supply a key from Settings → Agents through its secret configuration.
Add the workflow skill. It teaches source handling, monitor review and delivery checks. Install it in a supported client; the skill complements authentication.
Check access, then ask for news. Start with account.overview, check the returned account and plan, then search a defined topic and time window. News and monitoring need paid plan access.

https://www.sentrydock.com/mcp

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

pnpm add -g https://www.sentrydock.com/downloads/sentrydock-1.1.1.tgz
export SENTRYDOCK_API_KEY=YOUR_API_KEY
sentrydock news search "news in Peru"

curl -s https://www.sentrydock.com/api/v1/news/search \
  -H "Authorization: Bearer $SENTRYDOCK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query":"news in Peru","hours_ago":48,"limit":20}'

MCP server guide · CLI guide · API reference

What can the agent do?

Company watch

Follow earnings, regulatory actions and operational events with exact company identities and source links.

Market briefing

Turn commodity, energy and geopolitical coverage into a dated note the desk can check.

Policy and geopolitical watch

Keep official announcements, enacted rules and local reporting distinct across languages.

Alerts for a persistent agent

Send signed events to your own HTTPS receiver, inspect delivery and recover missed alerts.

Try a source-linked research task

Search the last 48 hours for LNG terminal disruptions. Group by facility, include original links and available publication times, and separate reported outages from speculation. Tell me where the indexed results do not support a conclusion. Do not change monitors.

Review the returned links and dates. Then ask for a monitor proposal with the sources, materiality rules, cadence and destination you want. Saving the monitor, completing a run and delivering an alert are separate results.

Recurring coverage, after the chat

SentryDock runs the saved monitor independently of your assistant. New monitors default to hourly; inspect the persisted schedule and current plan limits before promising realtime coverage. Ask the agent for recent alerts when you need them, or connect a signed callback for a persistent receiver.

“Keep watch and webhook me when something matters.”

Watch a customer’s announcements, a supplier’s official updates or the facts behind your research. Give the monitor clear triggers, source URLs and a schedule.

Create or choose a recurring monitor in your account.
Get a public HTTPS callback from your agent host or workflow. Use a destination you own or are authorized to send to.
Call delivery.connect below, or open Settings → Connections → Agent webhooks.
Save the returned signing secret securely. In Settings, send a signed test and check that your receiver accepts it.

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

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

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"}'

Replace MONITOR_ID with your monitor’s ID. Omit it to receive alerts from all your owned recurring monitors, including new ones. Team monitors require the team’s webhook permission.

Your receiver gets signed JSON with summaries, original source links and a stable event ID. Temporary delivery failures retry automatically. Read delivery.get or Settings for delivery status; use alerts.list to recover missed results.

A callback reaches your agent’s host; the host decides what to run. Connecting OAuth alone does not wake a chat assistant. No event is sent when a run has no relevant results.

Read the payload, signature and recovery guide

Tools and source coverage

account.overview: Your plan, access, monitors, delivery and latest alerts.
news.search / news.latest: Find source-linked items already in the news index.
news.mine / alerts.list: Read the hits from your own monitoring.
monitors.list / monitors.create: Inspect existing coverage and save a reviewed monitor.
monitors.run / monitors.runs / monitors.results: Run an authorised check and inspect execution and output.
delivery.connect: Connect a destination you chose, including signed webhooks.

Inspect the live schema before calling a tool. The account gateway does not change billing, invite a team, run predictions or delete your user.

Build coverage from websites and news, RSS, local publications, government sources, X, Telegram, Reddit and Truth Social. Access and cadence vary by source. Alerts can use your account delivery or an authorised Slack, Teams, Discord or webhook destination.

Frequently asked questions

Does the agent search the whole web live?: News search ranks items already indexed by SentryDock. An empty result does not establish that an event did not happen. Check the time window and source coverage, or set up a monitor for the sources you need.
Does monitoring continue after the chat ends?: Yes. A saved monitor runs in your SentryDock account on its configured schedule. The assistant reads alerts when it asks; a persistent agent needs a reachable signed webhook receiver for push delivery.
How quickly do alerts arrive?: The saved cadence and source access determine detection. New monitors default to hourly. Realtime availability and slots depend on the plan; a realtime option is not a guarantee of immediate detection from every source.
What does it cost?: The connection uses the same account and plan as the app. Overview and usage can be read before payment; indexed news and monitoring require paid access. Check the pricing page and account limits before running a larger coverage book.
Does the skill grant account access?: No. SKILL.md teaches the workflow. OAuth or a securely configured API key grants access. A local skill does not automatically become a cloud connector or an approved marketplace listing.
Can I control changes?: The agent acts on your account. Give it the exact monitor, source and destination scope, and review changes before authorising them. Remove the connector or revoke its API key to stop access; pause monitors separately to stop scheduled monitoring.

Plans and limits · Troubleshoot a monitor

Put your coverage in the conversation.

Connect, inspect your account and run one source-linked search.

Connect SentryDock