# core.blue (remote · mcp.core.blue)

Private, self-describing memory for agents: one house per customer, on a machine of its own.

- Trust score: 78/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- remote · `mcp.core.blue`: 78/100 (this document), [markdown](https://verifymcp.io/servers/blue-core-memory/mcp.md), [page](https://verifymcp.io/servers/blue-core-memory/mcp)

## Channel facts

- Endpoint: `https://mcp.core.blue/`
- Transports: `streamable-http`
- Auth: `none`
- Version: `0.2.1`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Endpoint Security**: 57/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - Authorisation check failed: no authorisation is required to call this server, and it exposes a tool marked destructive (close_sandbox).
  - HTTPS is enforced; there's no plaintext access path.
  - HSTS check failed: the Strict-Transport-Security header is absent.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 73/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 1617 tokens (~161/item across 10 items; 10 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 100/100
  - No destabilizing schema changes in the last 30 days.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 10 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 11 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the core.blue MCP server?

core.blue is a hosted endpoint at https://mcp.core.blue/, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http blue-core-memory 'https://mcp.core.blue/'
```

### Cursor

```json
{
  "mcpServers": {
    "blue-core-memory": {
      "url": "https://mcp.core.blue/"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "blue-core-memory": {
      "type": "http",
      "url": "https://mcp.core.blue/"
    }
  }
}
```

### Codex

```toml
[mcp_servers.blue-core-memory]
url = "https://mcp.core.blue/"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "blue-core-memory": {
      "type": "remote",
      "url": "https://mcp.core.blue/",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add blue-core-memory --url 'https://mcp.core.blue/' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  blue-core-memory:
    url: "https://mcp.core.blue/"
```

### Netclaw

```json
{
  "McpServers": {
    "blue-core-memory": {
      "Transport": "http",
      "Url": "https://mcp.core.blue/"
    }
  }
}
```

### Vellum

```bash
assistant mcp add blue-core-memory -t streamable-http -u 'https://mcp.core.blue/'
```

### Other

```json
{
  "mcpServers": {
    "blue-core-memory": {
      "type": "http",
      "url": "https://mcp.core.blue/"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-05 (score 78, 0)

- [security] Stability: 0.97 → pass

### 2026-11-04 (score 78, +1)

No change was recorded against any check on this day. Stability & Change Management went from 93 to 97. That category is still filling its 30-day observation window: 28 days of observed history at the previous scan, 29 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-02 (score 77, +1)

No change was recorded against any check on this day. Stability & Change Management went from 87 to 90. That category is still filling its 30-day observation window: 26 days of observed history at the previous scan, 27 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-31 (score 76, +1)

No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-29 (score 75, +1)

No change was recorded against any check on this day. Stability & Change Management went from 73 to 77. That category is still filling its 30-day observation window: 22 days of observed history at the previous scan, 23 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-27 (score 74, +1)

No change was recorded against any check on this day. Stability & Change Management went from 67 to 70. That category is still filling its 30-day observation window: 20 days of observed history at the previous scan, 21 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-25 (score 73, 0)

- [security] Tool “open_sandbox” rewrote its description, which is the text the model reads

### 2026-10-24 (score 73, +1)

No change was recorded against any check on this day. Stability & Change Management went from 57 to 60. That category is still filling its 30-day observation window: 17 days of observed history at the previous scan, 18 at this one. The score rises as the window fills, whether or not the server changes.

## MCP tools (10)

### `search_library` (~105 tokens)

Search the manual

Search the manual by your own words. Returns up to 'limit' topics ranked by relevance, each with slug, title, summary, tags and the slugs to read next. The manual is small: 'about' lists every topic, and you may not need a search at all.

Input parameters:

- `limit` (integer): Maximum number of hits, 1 to 20, default 5.
- `query` (string, required): What you want to know, as words or a question.

Output parameters:

- `hits` (array): The topics that match, best first.
- `note` (string): Empty when there are hits. Otherwise what to do instead.
- `topics_in_manual` (integer): How many topics the manual has in all.

### `open_sandbox` (~200 tokens)

Open a sandbox house

If you run in an assistant where a person adds connectors (ChatGPT, Claude.ai and others), do not call this: you could not enter the house it opens. Ask your person to add https://try.core.blue/ as a connector instead; it opens a house you can use. Otherwise: call this to try a house yourself before anyone decides. It opens a throwaway house for a few hours: a real house with the kernel of Atlantis, on core.blue's server, with no account, no key and no person involved. Returns the addresses of its doors (connect to 'modeller' with your MCP client and call overview), when it ends, and its limits. Whoever has the addresses can use the house, so keep them to yourself. It is deleted completely when it ends or when you call close_sandbox. If all places are taken, it says when the next one is free. One at a time from where you call.

Output parameters:

- `doors` (object|null): The addresses of the house's doors. Whoever has them can use the house. Null if nothing was opened.
- `ends_at` (string|null): When the house is stopped and deleted, UTC. Null if nothing was opened.
- `for_your_person` (string|null): Two sentences for the person you act for, to pass on as they are: how to try a house in an assistant where a person adds connectors (ChatGPT, Claude.ai and others), through https://try.core.blue/. Nu…
- `lease` (string|null): The lease: 'sb_' and 26 characters. Pass it to close_sandbox. Null if nothing was opened.
- `limits` (object): What a sandbox house may do and hold.
- `next_free_at` (string|null): When all places are taken: the time, UTC, at which the next one is free at the latest. Null otherwise.
- `note` (string): What happened and what to do next.
- `opened` (boolean): True if a house was opened for you.

### `about` (~55 tokens)

About core.blue

What core.blue is, in a form you can paste into a report: the product in one paragraph, what is live and what is not, what to call next, and every topic of the manual with a summary. Call this first.

Output parameters:

- `desk` (string): Whether request_house takes requests: 'closed' (it sends nothing and takes nothing) or 'waiting-list' (a person can confirm a request and is written to once when houses can be had; no house is built…
- `how_to_proceed` (array): What to call next, in order.
- `manual` (object): The version of the manual these answers come from; cite it in a report.
- `name` (string): The name of the product.
- `status` (string): What is live today and what is not.
- `summary` (string): What core.blue is, in one paragraph you can paste into a report.
- `topics` (array): Every topic of the manual in reading order, with a summary: choose from here and call read_topic.

### `pricing` (~106 tokens)

Prices

Call this when the person you act for asks what a house costs. The offer as data: what each size costs per month and whom it is meant for, the one-off price of the 14-day test, the price of additional volume, how model supply is charged, the terms, an example of the prices with tax, and what is never charged. 'undecided' lists the fields that are null because they are not decided yet; any other null means 'does not apply'.

Output parameters:

- `amounts_are` (string|null): 'net' if amounts exclude tax, 'gross' if they include it.
- `as_of` (string): The date this offer was last changed.
- `currency` (string|null): The currency of every amount, as an ISO code. Null if not decided.
- `gross_example` (object|null): What a customer pays with tax, for one example: whom it is for, the rate, and each size with its prices after tax. An example, not an invoice. Null if the offer gives none.
- `manual` (object): The version of the manual; cite it in a report.
- `model_supply` (array): The ways a house can get a language model, and how each is charged.
- `never_charged` (array): What is never charged for.
- `note` (string): What to keep in mind when quoting from this answer.
- `operator` (string): Who operates core.blue and sends the invoice.
- `regions` (object): Where the machines stand.
- `service_fee` (number): The share added to the price of everything bought from a supplier and passed on (volume, tokens): 0.2 means 20 percent.
- `sizes` (array): The sizes of a house.
- `source` (string): 'manual' if this is the offer of the manual, 'operator' if the operator of this server supplied its own.
- `stages` (array): The stages from first try to a lasting house.
- `terms` (object|null): The commercial terms. Null if none are stated.
- `undecided` (array): The paths of all fields that are null because they are not decided yet. A null field that is not listed here does not apply.
- `volume` (object): Additional data volume attached to a VM house.

### `leave_feedback` (~103 tokens)

Leave feedback

Tell the operator how it went, before you leave. Four questions, each optional, at least one answered, each at most 2000 characters. It is read by people at the next evaluation and deleted afterwards; send nothing secret.

Input parameters:

- `better` (string|null): What could we do better?
- `bothered` (string|null): What bothered you?
- `missing` (string|null): Which features are missing?
- `noticed` (string|null): What did you notice?

Output parameters:

- `note` (string): What happens with it.
- `recorded` (boolean): True if the feedback was taken.

### `request_house` (~223 tokens)

Request a house

Ask for a house on its own VM for the person you act for. No house can be had yet: while about says 'desk: waiting-list', this puts the person on the waiting list, builds nothing and costs nothing; while it says 'desk: closed', it answers 'accepted: false' and sends nothing. This sends an e-mail to a person: call it only when they know it is coming. They receive one mail with one link and confirm with a button on the page behind it; nothing happens without that click. Returns a ticket to follow with house_status.

Input parameters:

- `email` (string, required): E-mail address of the person who will confirm and own the house. One address, name@domain.tld.
- `language` (string): Language of the mail and of the page the person confirms on: 'en' or 'de'. Default 'en'.
- `size` (string): Size of the machine: 'economy' (small VM) or 'premium' (larger VM). Both run the same software. Default 'economy'.

Output parameters:

- `accepted` (boolean): True if the request was taken and a ticket exists. It does not say that a mail went out: 'note' says under which conditions one does.
- `note` (string): What happened and what to do next.
- `ticket` (object|null): The ticket of the request. Null if it was not accepted.

### `recommend` (~343 tokens)

Recommend a house

Call this once you know what the person you act for needs: whether a house fits, which size, what it costs per month, and a report in Markdown you can pass on. It advises against a house where one is the wrong choice. Name every need that applies, from this vocabulary. For a house: work_outlives_the_agent, corrections_must_stay_visible, sources_matter, own_vocabulary, confidential_material. Against: only_user_preferences_across_chats, only_retrieval_over_text, millisecond_answers_at_high_volume, nobody_will_model. Deciding the size: data_must_stay_in_own_network, scanned_documents, names_found_automatically, large_material. Asking for something planned: material_is_files, house_judges_while_away. Called without needs, it returns the vocabulary with a sentence per need. With language 'de' the report for the person is German; every other field stays English.

Input parameters:

- `data_gb` (number|null): How much material there is, in gigabytes, if you know. Used to calculate additional volume.
- `in_your_words` (string|null): Optional: what the person asked you for, in one or two sentences of your own. It is recorded so that we learn which questions we did not foresee. It is not interpreted and does not change the answer.
- `language` (string): Language of 'report', the text for the person: 'en' or 'de'. Default 'en'. Every other field stays English.
- `needs` (array|null): The needs that apply, as words of the vocabulary. Omit to get the vocabulary.

Output parameters:

- `cites` (array): The topics this answer rests on, each with the hash of its body.
- `gaps` (array): What you asked for that is not available yet, the most important first.
- `manual` (object): The version of the manual; the report cites it.
- `next_steps` (array): What to call next.
- `price` (object|null): What the recommended house costs. Null when a house does not fit or no need was named.
- `reasons` (array): One entry per need you named: what it means for a house, and the topic that backs it.
- `report` (string): The report for the person you act for, in Markdown, ready to pass on. Empty when no need was named.
- `size` (string|null): The size of house that follows from the needs today. A need for something planned does not change it; its gap says which size it will need. Null when a house does not fit or no need was named.
- `verdict` (string): 'fits', 'fits_with_gaps', 'does_not_fit', or 'say_what_you_need' when no need was named.
- `vocabulary` (array): The needs you can name. Filled only when you named none.

### `read_topic` (~72 tokens)

Read a topic

Read one topic of the manual in full by its slug (from about or search_library). The body is Markdown; 'see_also' lists the slugs to read next; 'sha256' proves what you read.

Input parameters:

- `slug` (string, required): The topic's slug, e.g. 'getting-a-house'.

Output parameters:

- `body` (string): The topic in Markdown.
- `see_also` (array): Slugs of the topics to read next.
- `sha256` (string): SHA-256 of the body as served, lower-case hex; cite it to prove what you read.
- `slug` (string): The stable identifier of the topic.
- `status` (string): 'draft' or 'published'. A draft has not been reviewed by the operator yet.
- `summary` (string): One or two sentences that answer the question the title asks.
- `tags` (array): Keywords; the first names the group of the topic.
- `title` (string): The title of the topic.

### `house_status` (~79 tokens)

Status of a house request

The state of a house request or house by its ticket id: requested, confirmed, expired, declined; later building, trial, licensed, dormant, exported, deleted. Says what to do next, and the address once the house exists. Ask no more than once an hour.

Input parameters:

- `ticket_id` (string, required): The ticket id returned by request_house.

Output parameters:

- `address` (string|null): The address of the house once it exists. Null before.
- `id` (string): The ticket id; pass it to house_status.
- `note` (string): What to do next.
- `state` (string): One of: requested (waits for the person's click), confirmed (the person clicked), expired (not confirmed in time), declined (the person said no); later: building, trial, licensed, dormant, exported,…

### `close_sandbox` (~135 tokens)

Close the sandbox house

End your sandbox now: the house and everything in it is deleted, and the place is free for the next agent. Please answer the four questions of leave_feedback here as well, each optional: what you noticed, what bothered you, what could be better, what is missing.

Input parameters:

- `better` (string|null): What could we do better?
- `bothered` (string|null): What bothered you?
- `lease` (string, required): The lease that open_sandbox returned: 'sb_' and 26 characters.
- `missing` (string|null): Which features are missing?
- `noticed` (string|null): What did you notice?

Output parameters:

- `closed` (boolean): True if the house was ended and deleted.
- `feedback_recorded` (boolean): True if feedback came with the call and was recorded.
- `note` (string): What happened.

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/blue-core-memory/mcp#diagnostics

## Score history

- 2026-11-05: 78
- 2026-11-04: 78
- 2026-11-03: 77
- 2026-11-02: 77
- 2026-11-01: 76
- 2026-10-31: 76
- 2026-10-30: 75
- 2026-10-29: 75
- 2026-10-28: 74
- 2026-10-27: 74
- 2026-10-26: 73
- 2026-10-25: 73
- 2026-10-24: 73
- 2026-10-23: 72
- 2026-10-22: 72
- 2026-10-21: 71
- 2026-10-20: 71
- 2026-10-19: 70
- 2026-10-18: 70
- 2026-10-17: 69
- 2026-10-16: 69
- 2026-10-15: 68
- 2026-10-14: 68
- 2026-10-13: 67
- 2026-10-12: 67
- 2026-10-10: 66
- 2026-10-07: 65

## Common questions

### What is the core.blue MCP server?

core.blue is an MCP server listed in the public MCP registry as blue.core/memory. Private, self-describing memory for agents: one house per customer, on a machine of its own. This page covers its hosted endpoint (https://mcp.core.blue/).

### Is the core.blue MCP server safe to use?

core.blue scores 78 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the core.blue MCP server expose?

core.blue exposes 10 tools: search_library, open_sandbox, about, pricing, leave_feedback, and 5 more. Their descriptions and schemas cost roughly 1,421 tokens of context every time the server is loaded.

### Does the core.blue MCP server require authentication?

No. We connected to core.blue without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the core.blue MCP server still maintained?

core.blue is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

## Links

- Remote endpoint: https://mcp.core.blue/
- Website: https://core.blue/
- Changelog RSS feed: https://verifymcp.io/servers/blue-core-memory/mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/blue-core-memory/mcp.json
- HTML version of this page: https://verifymcp.io/servers/blue-core-memory/mcp