@pipeworx/sec-nmfp

US money market fund portfolio holdings, from SEC Form N-MFP. Keyless.

Part of Pipeworx — an MCP gateway connecting AI agents to 1576+ live data sources.

Tools

• nmfp_security_holders(issuer, period?, limit?) — which money market funds hold a given issuer's paper. Accepts an issuer name fragment or a CUSIP.
• nmfp_fund_holdings(fund, period?, limit?) — one fund's reported portfolio for a month, with a per-category position count.
• nmfp_periods() — which monthly reporting periods SEC has published (currently 50) and the date range each covers.

The inversion is the point of this pack

SEC publishes N-MFP filing-first: one filing per fund per month, portfolio inside. So "what does this fund hold" is the natural question and "which funds hold this issuer's paper" is not — that one requires reading every filing in the period. nmfp_security_holders does exactly that on each call, which is the reason to have the pack at all. When a bank or an issuer wobbles, the question people ask is who is exposed, and there is no filing that answers it.

Measured on the 2026-07-09→2026-08-07 period: 341 fund filings, 47,129 holdings. Federal Home Loan Bank alone appears in 129 funds across 6,489 positions.

Reading traps, surfaced in the responses

• Positions are counts, not dollars. N-MFP position value is not carried here, so positions_by_category says how many holdings sit in each category, not how much money does. A fund running many small repo positions against one large Treasury block reads repo-heavy by count and may not be by value. Both tools say this in reading_note.
• Only money market funds file N-MFP. A bond or equity fund will never appear, however large. The found: false path says that rather than implying the fund holds nothing.
• Issuer names are as the fund typed them — FEDERAL HOME LOAN BANK SYSTEM, not a normalized entity. Matching is substring, and the miss path tells the caller to shorten the fragment rather than concluding no fund holds it.

Shape

One ZIP per monthly period, ~11.9MB, carrying 23 TSVs. Three are read: NMFP_SUBMISSION (the fund and its filing), NMFP_SERIESLEVELINFO, NMFP_SCHPORTFOLIOSECURITIES (the holdings). The rest are skipped — including NMFP_COLLATERALISSUERS, which is 55MB raw on its own, five times the whole rest of the archive, and answers no question asked here.

Read live per call rather than kept in a mirror: one period is one fetch, and a money-market question is nearly always about the latest month.

Two file-format notes worth keeping, because both have cost time in this codebase:

• Sizes come from the ZIP central directory, not local headers. SEC sets the data-descriptor flag on some datasets, which zeroes the local-header sizes; a reader that trusts them extracts nothing.
• The filenames encode a date range (20260709-20260807_nmfp.zip) that is not derivable — the windows do not align to calendar months and shift each cycle. This is the fourth SEC dataset here with its own path convention, so the index page is read and the URL is never constructed.

Cadence

Money market funds file monthly; SEC posts a period a few days after it closes.

Source

SEC Form N-MFP data sets — https://www.sec.gov/data-research/sec-markets-data/dera-form-n-mfp-data-sets

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "sec-nmfp": {
      "url": "https://gateway.pipeworx.io/sec-nmfp/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/sec-nmfp/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1576+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

{
  "mcpServers": {
    "sec-nmfp": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-sec-nmfp"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-sec-nmfp

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Sec Nmfp data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT