Pushary: human-in-the-loop for AI agents MCP server
io.github.Pushary/pusharyPusharystreamable-httpv1.5.2Scanned
Human-in-the-loop for AI agents: ask a person and get approvals or answers by push notification.
6 tools: 5 can delete or cancel things and 1 only read.
- Reads private data · 0
- Reads text others write · 4
- Can send data out · 1
- 1 finding, 1 high
Tools listed without sign-in. Remote posture
With Anlyon
5 tools here can delete or cancel things. Govern them.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the Pushary: human-in-the-loop for AI agents MCP server can do
The Pushary: human-in-the-loop for AI agents MCP server is published by the GitHub account Pushary. It has 6 tools. It can delete or cancel things (5 tools: schedule_reminder, send_notification, ask_user, propose_scope and 1 more). 1 tool only reads.
4 tools bring in text other people can write (schedule_reminder, send_notification, ask_user and 1 more); 1 tool can send data out of your system (send_notification).
The fixed checks matched 1 finding, 1 of them high severity (sensitive-path). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in, and its tools changed 2 times since Anlyon Radar first saw them.
Last scanned Nov 9, 2026, version 1.5.2.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
6 tools
Deletes or cancels 5Claude Code: deny
ask_userRequest an unresolved decision or missing input from the user through Pushary.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "request": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Request an unresolved decision or missing input from the user through Pushary. Supports confirm, select and input questions. Delivery and waiting follow the account policy; a call waits at most 48 seconds and may return immediately. Returns the question state, answer when available, delivery information and handoff fields. answerUrl is optional. A pending question remains answerable after a poll ends. Sends a real question; it does not grant permission for other actions.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- title: Ask User a Question
cancel_questionRetract a pending question that is no longer needed or is moving to the current client.No findings
Why it is in this group
- The name "cancel_question" says "cancel": a word for removing or ending something.
- The server marks it destructive (destructiveHint: true).
Description, as the server wrote it Retract a pending question that is no longer needed or is moving to the current client. Returns cancelled:true only when a pending question was removed. False may mean it was already answered, expired, missing or unavailable. An unavailable state includes handoffAction:stop because cancellation could not safely inspect or fence the question. Does not retract an answer already recorded.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: true
- openWorldHint: false
- title: Cancel Pending Question
propose_scopeRequest agreement on an unresolved or user-requested boundary for this session.Reads text others write1 finding
Why it is in this group
- The description's first verb is "request": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- High
The input schema at properties.offLimitsPaths.description refers to "**/.env*" and ".env".
What this check looks forMatched text
inputSchema.properties.offLimitsPaths.description: "..."**/" needs a directory before it, so "**/.env*" is expanded for you to also cover a ro..."
Description, as the server wrote it Request agreement on an unresolved or user-requested boundary for this session. Sends a real scope question and returns ratified, answered, contract and enforcement information. Only supported hooks can enforce file paths on path-bearing tool calls; promises are recorded but not enforced. Empty enforces means no automatic boundary checking. A contract never overrides host permissions or authorizes a separate action. The call waits at most 48 seconds; a pending proposal remains answerable after the call returns.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- title: Propose Run Scope
schedule_reminderSchedule, list, retrieve or cancel a personal Pushary reminder for the authenticated user.Reads text others writeNo findings
Why it is in this group
- The name "schedule_reminder" says "schedule": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Schedule, list, retrieve or cancel a personal Pushary reminder for the authenticated user. Scheduled reminders can arrive after the agent exits. One-time only; up to 30 days away, 25 pending per user/site. Scheduling requires body and either inMinutes or at with an explicit UTC offset. requestId identifies one schedule across retries. reminderId retrieves its outcome; no arguments lists pending reminders; cancelReminderId cancels a pending reminder. Returns the scheduled time, state and any delivery warning. Delivery respects stop/mute and active membership. Does not schedule agent work or customer campaigns. API reference: https://pushary.com/docs/agents/reference/tools#schedule_reminder
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- title: Schedule a Personal Reminder
send_notificationSend a notification to connected devices for a requested update or a meaningful unattended result, such as a finished task.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "send_notification" says "send": a word for sending, and what is sent cannot be taken back.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "send_notification" says "send": it sends or publishes something to other people.
Description, as the server wrote it Send a notification to connected devices for a requested update or a meaningful unattended result, such as a finished task. Delivery follows the account policy and optional recipient filters. context adds a detail page. context.askQuestion is kept for older clients; ask questions with ask_user. Returns web/mobile delivery counts and a warning when no delivery channel is connected. Sends real notifications. API reference: https://pushary.com/docs/agents/reference/tools#send_notification
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- title: Send Push Notification
Only reads 1Show 1 read-only tool
wait_for_answerRead the answer or current state of an existing Pushary question by its correlationId.No findings
Why it is in this group
- The name "wait_for_answer" says "wait": a word for reading.
Description, as the server wrote it Read the answer or current state of an existing Pushary question by its correlationId. Waits up to timeoutMs, capped at 48 seconds; a poll that ends with status pending is not an answer, so keep waiting. Returns answered, the answer when available, status and handoff information. Only pending is a live unanswered question; cancelled, expired, missing and unavailable are terminal states. Does not send another question.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- title: Wait for User Answer
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.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write4
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out1
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 19:00 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 5
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://pushary.com/.well-known/oauth-protected-resource/api/mcp/mcp
- Authorization servers
- https://clerk.pushary.com
- Legacy HTTP+SSE transport
- yes
- Dynamic client registration
- yes
- Client ID metadata documents
- no
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 5 tools that change things (5 destructive). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://pushary.com/api/mcp/mcp: tools/list returned 6 tools to an anonymous client, including schedule_reminder (destructive), send_notification (destructive), ask_user (destructive), propose_scope (destructive), cancel_question (destructive)
Findings
Matches of fixed checks, with the text they matched.
- Highpropose_scope
The input schema at properties.offLimitsPaths.description refers to "**/.env*" and ".env".
What this check looks forMatched text
inputSchema.properties.offLimitsPaths.description: "..."**/" needs a directory before it, so "**/.env*" is expanded for you to also cover a ro..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools3 changed.~3 changedView diff
- Changed:ask_userDeletes or cancelsdescriptioninputSchema
- Changed:propose_scopeDeletes or cancelsdescriptioninputSchema
- Changed:wait_for_answerOnly readsdescriptioninputSchema
Tools6 changed.~6 changedView diff
- Changed:ask_userDeletes or cancelsdescriptioninputSchemaannotations
- Changed:cancel_questionDeletes or cancelsannotations
- Changed:propose_scopeDeletes or cancelsdescriptionannotations
- Changed:schedule_reminderDeletes or cancelsdescriptionannotations
- Changed:send_notificationDeletes or cancelsdescriptioninputSchemaannotations
- Changed:wait_for_answerOnly readsdescriptionannotations
Install
Pinned to version 1.5.2. Values in angle brackets are yours to fill in.
claude mcp add --transport http pushary https://pushary.com/api/mcp/mcp --header 'Authorization: <your-authorization>'claude mcp add --transport http pushary https://pushary.com/api/mcp/mcp --header 'Authorization: <your-authorization>'Run in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__pushary__wait_for_answer"
],
"ask": [],
"deny": [
"mcp__pushary__schedule_reminder",
"mcp__pushary__send_notification",
"mcp__pushary__ask_user",
"mcp__pushary__propose_scope",
"mcp__pushary__cancel_question"
]
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as pushary. Review it before you use it.
Capability tags behind 4 rules
- mcp__pushary__schedule_reminder untrusted-input
- mcp__pushary__send_notification untrusted-input, outbound
- mcp__pushary__ask_user untrusted-input
- mcp__pushary__propose_scope untrusted-input
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/io.github.Pushary/pushary/verify?hash=bad7caf7c6cfa1af7b03a4a2ef29ef2fa1f38025d3bb63948d4a0a609f5c58eb" | jq -e '.changed == false'Pinned hash bad7caf7c6cf, from Nov 9, 2026, 05:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.Pushary/pushary/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.Pushary/pushary)Source: MCP Registry · repository. Synced Oct 25, 2026, 16:17 UTC.
Pushary: human-in-the-loop for AI agents MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Pushary: human-in-the-loop for AI agents MCP server safe?
- No scan can promise that. What its definitions show: 6 tools, 5 of which can change, send or delete something, and 1 finding (1 high) from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the Pushary: human-in-the-loop for AI agents MCP server delete data?
- Yes. 5 tools can delete, cancel or revoke something: schedule_reminder, send_notification, ask_user, propose_scope and cancel_question.
Can the Pushary: human-in-the-loop for AI agents MCP server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the Pushary: human-in-the-loop for AI agents MCP server have?
- 6 tools: 5 destructive, 1 read. schedule_reminder, send_notification, ask_user, propose_scope, cancel_question and wait_for_answer.
Does the Pushary: human-in-the-loop for AI agents MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Pushary: human-in-the-loop for AI agents MCP server send data to outside services?
- 1 tool can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): send_notification.
Has the Pushary: human-in-the-loop for AI agents MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 9, 2026. Every change is on this page and in its Atom feed.