ValuefocusValuefocus
Get the app
Filings & SEC

SEC EDGAR Agent Skill

Give an agent direct, citation-friendly access to company submissions and filing history.

Source
U.S. SEC
Access
No API key

What this adds to your agent

Retrieve filing history by CIK

Locate 10-K, 10-Q and 8-K forms

Return primary-source filing URLs

Implementation statusGuide reviewed · 2026-07
Implementation status

No API key

Expected result

Retrieve filing history by CIK

Known failure mode

Require a descriptive User-Agent, throttle requests and make the agent cite the original filing.

Compatibility

U.S. SEC · filings

Build the tool server

server.py
from mcp.server.fastmcp import FastMCP
import requests

mcp = FastMCP("SEC EDGAR")
HEADERS = {"User-Agent": "Research Agent contact@example.com"}

@mcp.tool()
def recent_filings(cik: str, form: str = "10-K") -> list:
    """Return recent SEC filings for a zero-padded CIK."""
    url = f"https://data.sec.gov/submissions/CIK{cik}.json"
    data = requests.get(url, headers=HEADERS, timeout=20).json()
    recent = data["filings"]["recent"]
    return [{k: recent[k][i] for k in ["form", "filingDate", "accessionNumber", "primaryDocument"]}
            for i, value in enumerate(recent["form"]) if value == form][:10]

if __name__ == "__main__": mcp.run()

Examples and expected outcomes

Example 1

Compare risk-factor changes between two 10-K filings

  • Sourced evidence
  • Repeatable steps
  • Explicit uncertainty
Example 2

Extract guidance from recent 8-K filings

  • Sourced evidence
  • Repeatable steps
  • Explicit uncertainty
Example 3

Trace a metric to its primary filing

  • Sourced evidence
  • Repeatable steps
  • Explicit uncertainty

Access requirements

Keys and runtime

No API key

License and data rights

Review source and upstream data licenses

Rate limits

Use caching, backoff and bounded concurrency

Connect to your Agent

Claude config
{
  "mcpServers": {
    "sec-edgar-agent": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}
Prompt example

“Use SEC EDGAR Agent Skill to research AAPL. List sources and timestamps first, then separate facts, calculations, risks and missing data. Do not give personalized trading advice.”

When not to use it

  • Require a descriptive User-Agent, throttle requests and make the agent cite the original filing.

Connect it to your agent

Claude Desktop
{
  "mcpServers": {
    "sec-edgar-agent": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}
Prompt to test

“Use the connected tool to research AAPL. State the source and timestamp for every factual claim, separate facts from interpretation, and do not give personalized financial advice.”

Production checklist

  • Keep secrets outside prompts and source code
  • Allowlist read-only tools first
  • Log tool inputs, outputs and source timestamps
  • Require citations and handle missing data explicitly

When not to use it

Require a descriptive User-Agent, throttle requests and make the agent cite the original filing.

Open official documentation

Project-specific questions

What can SEC EDGAR Agent Skill do in a real workflow?

Retrieve filing history by CIK Locate 10-K, 10-Q and 8-K forms Return primary-source filing URLs

What access and compatibility does SEC EDGAR Agent Skill require?

No API key. U.S. SEC.

What is the main implementation risk with SEC EDGAR Agent Skill?

Require a descriptive User-Agent, throttle requests and make the agent cite the original filing.

Related agent tutorials