LMGram MCP server
io.github.saezbaldo/lmgramsaezbaldostreamable-httpv0.1.2Scanned
Consent-based human matchmaking for peer advice, collaboration, and practical help.
17 tools: 1 can delete or cancel things, 3 do something that can't be undone, 8 change things and 5 only read.
- Reads private data · 3
- Reads text others write · 14
- Can send data out · 4
- 3 findings
Tools listed without sign-in. Remote posture
With Anlyon
1 tool here can delete or cancel things and 3 do something that can't be undone. 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 LMGram MCP server can do
The LMGram MCP server is published by the GitHub account saezbaldo. It has 17 tools. It can delete or cancel things (1 tool: panel_dismiss_suggestion), do things that cannot be undone, such as sending, paying or publishing (send_message, panel_send_message and panel_create_intent) and create or change data (8 tools: request_human_match, start_chat, open_lmgram_panel, panel_register_device and 4 more). 5 tools only read.
3 tools read private data (list_messages, panel_snapshot and panel_chat); 14 tools bring in text other people can write (request_human_match, start_chat, list_messages and 11 more); 4 tools can send data out of your system (send_message, panel_mark_read, panel_sync_device_keys and 1 more). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.
The fixed checks matched 3 findings (credentials-mention and annotation-conflict). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 21, 2026, version 0.1.2.
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.
17 tools
Deletes or cancels 1Claude Code: deny
panel_dismiss_suggestionDecline an unopened match suggestion.Reads text others writeNo findings
Why it is in this group
- The name "panel_dismiss_suggestion" says "dismiss": 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 Decline an unopened match suggestion.
Annotations the server declares
- title: Dismiss suggestion
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: true
- idempotentHint: true
Can't be undone 3Claude Code: ask
panel_create_intentPublish only the minimized intent explicitly reviewed and submitted by the user in this conversation panel.Reads text others writeNo findings
Why it is in this group
- The name "panel_create_intent" says "create": a word for a change.
- The description's first verb is "publish": a word for a change that another call does not undo.
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 Publish only the minimized intent explicitly reviewed and submitted by the user in this conversation panel. Do not include secrets, private transcripts, or dating requests.
Annotations the server declares
- title: Publish approved intent
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: false
panel_send_messageSend a client-encrypted text message after the user presses Send.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "panel_send_message" says "send": a word for sending, and what is sent cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "panel_send_message" says "send": it sends or publishes something to other people.
- outboundThe input takes a destination (input.chatId): what the tool writes goes there.
Description, as the server wrote it Send a client-encrypted text message after the user presses Send. Retry only the identical payload and clientMessageId.
Annotations the server declares
- title: Send encrypted message
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
send_messageSends a message into a LMGram chat.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "send_message" says "send": a word for sending, and what is sent cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "send_message" says "send": it sends or publishes something to other people.
- outboundThe input takes a destination (input.chatId): what the tool writes goes there.
Description, as the server wrote it Sends a message into a LMGram chat.
Annotations the server declares
- title: Send a chat message
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: false
Changes things 8Claude Code: ask
open_lmgram_panelOpen LMGram chats, groups and suggestions for the current ChatGPT conversation.1 finding
Why it is in this group
- The name "open_lmgram_panel" says "open": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "open_lmgram_panel" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "open_lmgram_panel" says "open": a word for a change.
Description, as the server wrote it Open LMGram chats, groups and suggestions for the current ChatGPT conversation. Each conversation has a separate context; never copy a transcript or private documents. This does not publish an intent or send a message.
Annotations the server declares
- title: Open LMGram conversation panel
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
panel_join_groupAccept an existing group invitation related to this context.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
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 Accept an existing group invitation related to this context.
Annotations the server declares
- title: Join group
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
panel_mark_readMark the displayed chat as read.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "panel_mark_read" says "mark": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a destination (input.chatId): what the tool writes goes there.
Description, as the server wrote it Mark the displayed chat as read.
Annotations the server declares
- title: Mark chat read
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
panel_register_deviceRegister only a public encryption key for this user's panel.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
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 Register only a public encryption key for this user's panel.
Annotations the server declares
- title: Register encryption device
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
panel_start_chatRequest a chat with a suggestion in this context, without sending a plaintext introduction.Reads text others writeNo findings
Why it is in this group
- The name "panel_start_chat" says "start": a word for a change.
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 a chat with a suggestion in this context, without sending a plaintext introduction.
Annotations the server declares
- title: Request conversation
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
panel_sync_device_keysShare client-encrypted message keys only with this account's registered devices.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "panel_sync_device_keys" says "sync": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a destination (input.chatId): what the tool writes goes there.
Description, as the server wrote it Share client-encrypted message keys only with this account's registered devices. Does not send messages, plaintext or private keys.
Annotations the server declares
- title: Sync own encryption devices
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: true
request_human_matchCreates an opt-in LMGram human match request only after the user explicitly asks to use LMGram to find a collaborator, peer, mentor, or helper.Reads text others write2 findings
Why it is in this group
- The name "request_human_match" says "request": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...tic matchmaking. Never include secrets, credentials, API keys, raw transcripts, customer da..."
- Info
The input schema at properties.summary.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.summary.description: "...-sensitive summary. Never send secrets, credentials, API keys, raw transcripts, customer da..."
Description, as the server wrote it Creates an opt-in LMGram human match request only after the user explicitly asks to use LMGram to find a collaborator, peer, mentor, or helper. Never invoke for general advice, dating or romantic matchmaking. Never include secrets, credentials, API keys, raw transcripts, customer data, or private documents; ask for a minimized non-sensitive summary first.
Annotations the server declares
- title: Request a human match
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: false
start_chatStarts a LMGram chat for a match request, or records that this side wants to chat.Reads text others writeNo findings
Why it is in this group
- The name "start_chat" says "start": a word for a change.
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 Starts a LMGram chat for a match request, or records that this side wants to chat.
Annotations the server declares
- title: Start or request a chat
- readOnlyHint: false
- openWorldHint: true
- destructiveHint: false
- idempotentHint: false
Only reads 5Show 5 read-only tools
get_current_userReturns the LMGram profile linked to the connected account.No findings
Why it is in this group
- The name "get_current_user" says "get": a word for reading.
Description, as the server wrote it Returns the LMGram profile linked to the connected account.
Annotations the server declares
- title: Get current LMGram profile
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
list_match_requestsLists recent match requests for the current LMGram user, optionally filtered by role and status.No findings
Why it is in this group
- The name "list_match_requests" says "list": a word for reading.
Description, as the server wrote it Lists recent match requests for the current LMGram user, optionally filtered by role and status.
Annotations the server declares
- title: List LMGram match requests
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
list_messagesLists messages in a LMGram chat.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "list_messages" says "list": a word for reading.
What it can reach
- private-dataThe name "list_messages" says "list" and "message": it reads messages.
- private-dataThe description's first sentence says "list" and "message": it reads messages.
- untrusted-inputThe name "list_messages" says "list" and "message": it reads messages, which other people can write.
- untrusted-inputThe description's first sentence says "list" and "message": it reads messages, which other people can write.
Description, as the server wrote it Lists messages in a LMGram chat. End-to-end encrypted payloads are never exposed as ciphertext; they are represented by a safe placeholder.
Annotations the server declares
- title: List chat messages
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
panel_chatRetrieve authorized chat data for client-side decryption, not model context.Reads private dataReads text others writeNo findings
Why it is in this group
- The description's first verb is "retrieve": a word for reading.
What it can reach
- private-dataThe name "panel_chat" says "chat": it reads messages.
- private-dataThe description's first sentence says "retrieve" and "chat": it reads messages.
- untrusted-inputThe name "panel_chat" says "chat": it reads messages, which other people can write.
- untrusted-inputThe description's first sentence says "retrieve" and "chat": it reads messages, which other people can write.
Description, as the server wrote it Retrieve authorized chat data for client-side decryption, not model context.
Annotations the server declares
- title: Read private chat
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
panel_snapshotRefresh this conversation's private panel data.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "panel_snapshot" says "snapshot": a word for reading.
What it can reach
- private-dataThe description's first sentence says "conversation": it reads messages.
- untrusted-inputThe description's first sentence says "conversation": it reads messages, which other people can write.
Description, as the server wrote it Refresh this conversation's private panel data.
Annotations the server declares
- title: Refresh conversation
- readOnlyHint: true
- openWorldHint: false
- destructiveHint: false
- idempotentHint: true
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.
168 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 3 private-data tools breaks every one.
- Reads private data3
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write14
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out4
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__lmgram__list_messages",
"mcp__lmgram__panel_chat",
"mcp__lmgram__panel_snapshot"
]
}
}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, 19:10 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
- 12
- Protocol version
- 2025-03-26
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://api.lmgram.com/.well-known/oauth-protected-resource
- Authorization servers
- https://api.lmgram.com
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- yes
- Client ID metadata documents
- no
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 12 tools that change things (1 destructive, 3 irreversible, 8 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://api.lmgram.com/mcp: tools/list returned 17 tools to an anonymous client, including panel_dismiss_suggestion (destructive), send_message (irreversible), panel_send_message (irreversible), panel_create_intent (irreversible), request_human_match (write) and 7 more
Findings
Matches of fixed checks, with the text they matched.
The server marks "open_lmgram_panel" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "open_lmgram_panel" says "open": a word for a change.
The description mentions "credentials".
What this check looks forMatched text
description: "...tic matchmaking. Never include secrets, credentials, API keys, raw transcripts, customer da..."
The input schema at properties.summary.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.summary.description: "...-sensitive summary. Never send secrets, credentials, API keys, raw transcripts, customer da..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.
Install
Pinned to version 0.1.2. Values in angle brackets are yours to fill in.
claude mcp add --transport http lmgram https://api.lmgram.com/mcpclaude mcp add --transport http lmgram https://api.lmgram.com/mcpRun 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__lmgram__get_current_user",
"mcp__lmgram__list_match_requests",
"mcp__lmgram__list_messages",
"mcp__lmgram__panel_snapshot",
"mcp__lmgram__panel_chat"
],
"ask": [
"mcp__lmgram__request_human_match",
"mcp__lmgram__start_chat",
"mcp__lmgram__send_message",
"mcp__lmgram__open_lmgram_panel",
"mcp__lmgram__panel_register_device",
"mcp__lmgram__panel_mark_read",
"mcp__lmgram__panel_sync_device_keys",
"mcp__lmgram__panel_send_message",
"mcp__lmgram__panel_start_chat",
"mcp__lmgram__panel_join_group",
"mcp__lmgram__panel_create_intent"
],
"deny": [
"mcp__lmgram__panel_dismiss_suggestion"
]
}
}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 lmgram. Review it before you use it.
Capability tags behind 14 rules
- mcp__lmgram__request_human_match untrusted-input
- mcp__lmgram__start_chat untrusted-input
- mcp__lmgram__list_messages private-data, untrusted-input
- mcp__lmgram__send_message untrusted-input, outbound
- mcp__lmgram__panel_snapshot private-data, untrusted-input
- mcp__lmgram__panel_chat private-data, untrusted-input
- mcp__lmgram__panel_register_device untrusted-input
- mcp__lmgram__panel_mark_read untrusted-input, outbound
- mcp__lmgram__panel_sync_device_keys untrusted-input, outbound
- mcp__lmgram__panel_send_message untrusted-input, outbound
- mcp__lmgram__panel_start_chat untrusted-input
- mcp__lmgram__panel_dismiss_suggestion untrusted-input
- mcp__lmgram__panel_join_group untrusted-input
- mcp__lmgram__panel_create_intent 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.saezbaldo/lmgram/verify?hash=bd577e126fac4d5c66ce01025961c19738612233688bafac08f6117eeaeb82e4" | jq -e '.changed == false'Pinned hash bd577e126fac, from Nov 21, 2026, 03:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.saezbaldo/lmgram/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.saezbaldo/lmgram)Source: MCP Registry. Synced Oct 21, 2026, 18:48 UTC.
LMGram MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the LMGram MCP server safe?
- No scan can promise that. What its definitions show: 17 tools, 12 of which can change, send or delete something, and 3 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the LMGram MCP server delete data?
- Yes. 1 tool can delete, cancel or revoke something: panel_dismiss_suggestion.
Can the LMGram MCP server do things that cannot be undone?
- Yes. 3 tools can send, pay, publish or merge: send_message, panel_send_message and panel_create_intent.
What tools does the LMGram MCP server have?
- 17 tools: 1 destructive, 3 irreversible, 8 write, 5 read. panel_dismiss_suggestion, send_message, panel_send_message, panel_create_intent, request_human_match, start_chat, open_lmgram_panel, panel_register_device and 9 more.
Does the LMGram MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the LMGram MCP server send data to outside services?
- 4 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): send_message, panel_mark_read, panel_sync_device_keys and panel_send_message.
Has the LMGram MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 21, 2026.