# Spekkr (remote · spekkr.com)

Public registry of specs AI agents can build from. Search, pull and publish over MCP.

- Trust score: 65/100 (medium)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `spekkr.com`: 65/100 (this document), [markdown](https://verifymcp.io/servers/com-spekkr-spekkr/spekkr.md), [page](https://verifymcp.io/servers/com-spekkr-spekkr/spekkr)

## Channel facts

- Endpoint: `https://spekkr.com/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `0.2.0`

## 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 2026-11-21.

- **Endpoint Security**: 63/100
  - 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 (unpublish_spec).
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - 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**: 67/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 1581 tokens (~197/item across 8 items; 8 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 0/100
  - Stability check failed: schema churn in the 4 days we've observed: 2 tool removals, 0 breaking changes, 0 auth/transport breaks, 0 additions.
- **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.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - All 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.
  - An AI judge read all 9 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the Spekkr MCP server?

Spekkr is a hosted endpoint at https://spekkr.com/mcp, 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 com-spekkr-spekkr 'https://spekkr.com/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "com-spekkr-spekkr": {
      "url": "https://spekkr.com/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "com-spekkr-spekkr": {
      "type": "http",
      "url": "https://spekkr.com/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.com-spekkr-spekkr]
url = "https://spekkr.com/mcp"
```

### opencode

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

### OpenClaw

```bash
openclaw mcp add com-spekkr-spekkr --url 'https://spekkr.com/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  com-spekkr-spekkr:
    url: "https://spekkr.com/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "com-spekkr-spekkr": {
      "Transport": "http",
      "Url": "https://spekkr.com/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add com-spekkr-spekkr -t streamable-http -u 'https://spekkr.com/mcp'
```

### Other

```json
{
  "mcpServers": {
    "com-spekkr-spekkr": {
      "type": "http",
      "url": "https://spekkr.com/mcp"
    }
  }
}
```

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.

### 2026-11-18 (score 65, 0)

- [security regression] Stability: unverified → fail
- [security regression] A breaking change shipped without a version bump: still 0.2.0
- [security regression] Tool “get_bounty” was removed
- [security regression] Tool “list_bounties” was removed
- [functional regression] Schema quality: 173 → 197

### 2026-11-17 (score 65)

First indexed and scored.

## MCP tools (8)

### `search_specs` (~327 tokens)

Search specs

Finds specs on Spekkr by words, the best match first, with each one's id, summary and platforms. Most build a part of an app; a spec that builds something else, such as a CAD model, a circuit or an animation, says so under kind. Search with a few words that name the thing, such as "sign in" or "sliding puzzle"; a slip in a long word is forgiven. When no spec holds every word, the specs that hold some of them are returned and partial is true. Leave query out to list the specs, the ones pulled most first. An answer holds at most limit specs: total counts the specs that matched, and offset reads further into them. catalogTotal counts every spec there is. A spec with publishedBy was published by a member.

Input parameters:

- `kind` (string): Only specs of this kind, such as CAD or animation. Each spec names its kind in its own words, so any text is allowed; capitals are ignored. A spec that names no kind builds software.
- `limit` (integer): How many specs to return at most. 20 when left out.
- `offset` (integer): How many of the specs that matched to pass over first. 0 when left out.
- `platform` (string): Only specs that target this platform.
- `query` (string): Words to look for in a spec's name, keywords, id, kind, stack, summary and platforms. A word may be only the start of one. Small words such as "the" and "for" are left out.

### `get_spec` (~121 tokens)

Read a spec's listing

Returns everything Spekkr knows about one spec: its manifest (what it builds, what the person may change under customize, the files, what it needs, the commands and secrets it uses under safety, the changelog). Read it before pull_spec to learn which choices the spec offers. Where the site has found specs like this one, related lists a few: each is another spec, with its own id to call get_spec with, and none of them is part of this one.

Input parameters:

- `id` (string, required): The spec's id, like username/name.

### `pull_spec` (~181 tokens)

Pull a spec to build from

Returns a spec's files, ready to write into a folder and build from: BUILD.md (how to build), spekkr.json (the manifest), spekkr.inputs.json (the person's choices) and the spec files. It checks the choices first and refuses a key the spec does not declare or a value it does not offer. Anything not chosen keeps the spec's default.

Input parameters:

- `choices` (object): The person's choices, by key from manifest.customize: any text for a setting, true or false for an option (on and off are taken too), one of its choices for a swap. For example {"database": "sqlite",…
- `id` (string, required): The spec's id, like username/name.

### `get_spec_file` (~113 tokens)

Read one spec file

Returns one file of a spec as text. Use it for a file pull_spec listed without content, unless its note says to download it. A long file comes in parts: when the answer has nextLine, call again with that line.

Input parameters:

- `id` (string, required): The spec's id, like username/name.
- `line` (integer): The first line to return. Leave it out to start at the top.
- `path` (string, required): The file's path, as pull_spec and get_spec list it.

### `check_spec` (~72 tokens)

Check a spec before listing it

Checks a spec someone is writing against Spekkr's listing rules and returns every problem at once. Pass the text of its spekkr.json. Nothing is saved or published. When it is ready, publish_spec publishes it.

Input parameters:

- `manifest` (string, required): The whole text of the spec's spekkr.json.

### `whoami` (~68 tokens)

Who am I on Spekkr

Says which Spekkr account the caller is signed in as and which specs it has published. This tool needs sign-in: the first call opens a Spekkr page in the browser, where the person signs in with a code sent to their email address and picks a username if they are new.

### `publish_spec` (~346 tokens)

Publish a spec

Publishes a spec on Spekkr under the signed-in person's username, as username/name, or publishes a new version of one of theirs. Send the text of spekkr.json and every file it lists. A small text file goes in files with its content. Any other file (an image, a PDF, a video, a long document) goes in files with its size in bytes and no content: the answer then gives an address to upload those files to with HTTP PUT, and once they are uploaded you call publish_spec again with the same arguments. Nothing is listed until a call answers with published. A listed spec names the member who published it. A published version never changes: to update a spec, raise the version, add a changelog entry and list every file again. A file that has not changed goes in files with keep set to true and nothing else, and Spekkr takes it from the version listed now, so a screenshot need not be uploaded again. In this version a published spec is free. If anything has to change first, the answer names every problem. This tool needs sign-in. Ask the person before you call it: publishing makes the files public.

Input parameters:

- `files` (array, required): Every file the manifest lists, every file inside a folder it lists, and every screenshot under demo.screenshots. Leave out spekkr.json, BUILD.md and spekkr.inputs.json. Each file has a path and one o…
- `manifest` (string, required): The whole text of the spec's spekkr.json. Its id is username/name, with the signed-in person's username, and its author.id is that username.

### `unpublish_spec` (~79 tokens)

Take a spec down

Takes one of the signed-in person's specs off Spekkr: its page, its files and its place in search. The versions it had stay used up, so publishing it again needs a higher version. This tool needs sign-in. Ask the person before you call it.

Input parameters:

- `id` (string, required): The spec's id, like username/name.

## Diagnostics

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

## Score history

- 2026-11-21: 65
- 2026-11-20: 65
- 2026-11-19: 65
- 2026-11-18: 65
- 2026-11-17: 65

## Common questions

### What is the Spekkr MCP server?

Spekkr is an MCP server listed in the public MCP registry as com.spekkr/spekkr. Public registry of specs AI agents can build from. Search, pull and publish over MCP. This page covers its hosted endpoint (https://spekkr.com/mcp).

### Is the Spekkr MCP server safe to use?

Spekkr scores 65 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 Spekkr MCP server expose?

Spekkr exposes 8 tools: search_specs, get_spec, pull_spec, get_spec_file, check_spec, and 3 more. Their descriptions and schemas cost roughly 1,307 tokens of context every time the server is loaded.

### Does the Spekkr MCP server require authentication?

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

### Is the Spekkr MCP server still maintained?

Spekkr is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What did VerifyMCP test on the Spekkr MCP server?

VerifyMCP ran 13 checks on Spekkr, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- Remote endpoint: https://spekkr.com/mcp
- Website: https://spekkr.com/
- Changelog RSS feed: https://verifymcp.io/servers/com-spekkr-spekkr/spekkr.xml
- Changelog JSON feed: https://verifymcp.io/servers/com-spekkr-spekkr/spekkr/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/com-spekkr-spekkr/spekkr.json
- HTML version of this page: https://verifymcp.io/servers/com-spekkr-spekkr/spekkr
