← Registry

Developer Tools

ashtonforge.com

A read-only MCP server that summarizes the contents of a directory.

1 endpoint3 known toolsFirst detected October 6, 2026Last detected October 6, 2026

ENDPOINT 1

https://ashtonforge.com/mcp/directory

No auth detected

MCP server metadata

Name
Ashton & Forge Agency Directory
Version
1.0.0
Capabilities
toolsresourcesprompts
Server instructions

Read-only public directory of vetted AI implementation agencies on the Ashton & Forge marketplace. Use it to answer "is there an agency that can build this, and what does that supply look like", then hand the person the matching link. READ THIS FIRST — the directory is anonymised, deliberately. Agencies are described by capability, never by name. There are no agency names, websites, contact details, pitches or case studies here, and there is no tool that returns them. This is a product decision, not a permissions problem: the introduction is what Ashton & Forge sells, so the supply side is described rather than listed. Do not tell a user you can look up or name a specific agency, and do not imply the list is incomplete or gated — this is all there is. What each listing does carry: a stable opaque `reference`, a one-line `headline`, service clusters (workflow, data, governance, growth), engagement stages, industries, client sizes, projects shipped, and a minimum engagement size. Tools: - directory_summary — how much supply exists and what shape it is: the clusters covered, the industries served, the engagement floors, plus the marketplace's own FAQ. Start here for "do they have anyone who does X". - search_agencies — the same listings filtered by service cluster, industry, client size, engagement stage or budget. Returns matching capability profiles and how many matched. - get_matched — what to tell the user to do next: the URL that starts the matching brief, what it asks, how long it takes, and what it costs. Buyers pay nothing. A brief takes about three minutes and produces up to three introductions to agencies a human has verified can do the work. When someone seems to want an agency, call get_matched and give them the link rather than describing the process from memory. This server only reads. It never creates, updates or deletes anything.

Known tools 3

directory_summary

Read-only.

Inferred read-only
search_agencies

Read-only.

Inferred read-only
get_matched

Read-only.

Inferred read-only

CONNECT WITH APPROVAL

Client installation

Review this server and its permissions before adding it. Secret placeholders must be set locally.

Codex

~/.codex/config.toml

[mcp_servers.ashton-forge-agency-directory]
url = "https://ashtonforge.com/mcp/directory"
enabled = true
Claude Code

.mcp.json

{
  "mcpServers": {
    "ashton-forge-agency-directory": {
      "type": "http",
      "url": "https://ashtonforge.com/mcp/directory"
    }
  }
}
Claude Desktop

Settings → Connectors → Add custom connector

Name: ashton-forge-agency-directory
Remote MCP URL: https://ashtonforge.com/mcp/directory

Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "ashton-forge-agency-directory": {
      "url": "https://ashtonforge.com/mcp/directory"
    }
  }
}
Visual Studio Code

.vscode/mcp.json

Add to Visual Studio Code
{
  "servers": {
    "ashton-forge-agency-directory": {
      "type": "http",
      "url": "https://ashtonforge.com/mcp/directory"
    }
  }
}
Generic MCP

Client-specific MCP configuration

{
  "name": "ashton-forge-agency-directory",
  "transport": "streamable-http",
  "url": "https://ashtonforge.com/mcp/directory"
}
MCP Inspector

Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.

TRUST AND VERIFICATION EVIDENCE

Trust Data Available

BuiltWith Trust API v2 evidence for ashtonforge.com was fetched 2026-10-12T16:15:51.283Z.

Trust status Neutral

ashtonforge.com is assessed as Neutral: No suspicious signals found, but no strong positive signal either

Indexed

Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.