Model Context Protocol (MCP) Server for OpenCTI
Project description
An MCP server front-end for pycti.
Inspired by Spathodea-Network/opencti-mcp, but rather than trying to reflect OpenCTI content back to the caller, the aim here is to condense, normalize, and consolidate data from OpenCTI into JSON for the LLM to consume in order to acheive the following improvements:
- More verbosity in field naming so the LLM can better intuit what a field represents
- Resolve parts of GraphQL-linked entities into the response so more context is available in a single MCP response
- Reduce the inclusion of non-informative metadata to reduce context window usage
To use:
- Set
$OPENCTI_URLand$OPENCTI_KEYto your OpenCTI URL and API Key, respectively. Or, you can provide these on the command-line. - Run
uvx pycti-mcp@latest
Usage details:
usage: pycti-mcp [-h] [-p PORT] [-s] [-v] [-u URL] [-k KEY]
Execute the OpenCTI MCP Server
options:
-h, --help show this help message and exit
-p, --port PORT TCP port to listen on (default 8002 - only used if -s/--sse is provided)
-s, --sse Start an SSE server (default: off)
-v, --verbose Run in VERBOSE mode (INFO level logging). Default: off (WARN level logging)
-u, --url URL OpenCTI URL - Can also be provided in OPENCTI_URL environment variable
-k, --key KEY OpenCTI API Key - Can also be provided in OPENCTI_KEY environment variable
Adding New Tools
New tools may be added by creating a new Python module under src/pycti_mcp/pycti_tools/, then adding it to the
the __all__ list in pycti_tools/__init__.py. The only requirement is that your code
must implement exactly one tool per module, and the module must contain a class ToolSpec that describes the
tool using the following three class variables:
name: The name of the tool as it will be exposed by the MCP serverdescription: A description of the tool to provide via the MCP server as willfn: The function that will act as the entrypoint for the toolopencti_url/opencti_key: Set to"", will be overwritten at run-time with the provided values
Example from lookup_reports.py below:
class ToolSpec:
name = "opencti_reports_lookup"
description = """Search in OpenCTI for any reports matching the search term `search`, and having creation timestamps between
`earliest` and `latest`. Any of these input variables can be omitted by setting them to None, if the aren't
desired for filtering reports. The result will be a list of structured objects representing all the reports
matching the provided criteria.
"""
fn = opencti_reports_lookup
opencti_url = ""
opencti_key = ""
Implemented Tools
OpenCTI Observable Lookup
Name: opencti_observable_lookup
Inputs: observable (str): An Observable
This tool will perform an exact-match lookup in OpenCTI for the observable value provided as observable.
Given an observable, queries for it in OpenCTI and, if it exists, returns JSON object representing the findings from OpenCTI for the observable, with the following fields:
observable_value: The observable value, as it is recorded in OpenCTIstix_id: The STIX Id of the observable objectopencti_id: The entity Id of the observable object in OpenCTIdata_type: The STIX Observable typedescriotion: A short description of the observable, from OpenCTIcreated: Creation data within OpenCTIlast_updated: The last time an update was written to the observable object in OpenCTIlabels: A list of labels (as strings) attached to the observableexternal_reports: A list of external reports containing the observablename: The title of the reporturls: List of URLs to fetch the report (or parts of it)
notes: Notes in OpenCTI written about the observableopinions: Opinions in OpenCTI about the observablesentiment: The sentiment expressed in the opinion.explanation: An explanation of the opinion.
OpenCTI Adversary Lookup
Name: opencti_adversary_lookup
Inputs: name (str): A name or alias of an adversary, intrusion set, threat actor, threat group, or campaign
This tool will search across all "adversary" type entities: Intrusion Sets, Actors, and Campaigns for the adversary
matching name either in its formal name or one of its aliases.
stix_id: The STIX ID of the adversary object.opencti_id: The entity ID of the adversary object in OpenCTI.name: The name of the adversary.data_type: The type of the entity (e.g., "Threat Actor").description: A brief description of the adversary.created: The creation date of the adversary in OpenCTI.last_updated: The last time the adversary was updated in OpenCTI.labels: A list of labels (as strings) attached to the adversary.first_seen: The first date the adversary was observed.last_seen: The last date the adversary was observed.external_reports: A list of external reports related to the adversary, each containing:name: The title of the report.urls: List of URLs to access the report or its parts.
notes: A collection of notes associated with the adversary.opinions: A list of opinions about the adversary, where each opinion includes:sentiment: The sentiment expressed in the opinion.explanation: An explanation of the opinion.
OpenCTI Report Lookup
Name: opencti_report_lookup
Inputs:
search(str): An optional search term to use to filter to reports matching a string termearliest(str): Optional timestamp that sets the earliest date to search for reportslatest(str): Optional timestamp that sets the latest date to search for reports
This tool will perform a lookup in OpenCTI of all of the threat reports matching a search term provided as search,
between the creation timestamps earliest and latest. Any of the inputs can be omitted (specified as None).
stix_id: The STIX ID of the report.opencti_id: The entity ID of the report in OpenCTI.name: The name of the report.data_type: The type of the entity (e.g., "Report").description: A brief description of the contents of the report.created: The creation date of the report.modified: The most recent modification date of the report.published: The report's publication date.labels: A list of labels (as strings) attached to the report.external_urls: A list of external URLs referencing sourcing of the report.report_types: The type label(s) of the analysis report.objects: The STIX objects (Entities and Cyber observables) contained within the report.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file pycti_mcp-0.1.3.tar.gz.
File metadata
- Download URL: pycti_mcp-0.1.3.tar.gz
- Upload date:
- Size: 8.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via: uv/0.7.19
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
91e3e85527db00a935574a5258e9ca0692ce1212dbcfea1b7cecab7c48b8bb3f
|
|
| MD5 |
0b85328c67bd6bd51f32f0fbd4dd1beb
|
|
| BLAKE2b-256 |
cc0ad1b00b11ff6986e1c3a767d35ee076b43d7abc8cb12e10f438a9694f4ee6
|
File details
Details for the file pycti_mcp-0.1.3-py3-none-any.whl.
File metadata
- Download URL: pycti_mcp-0.1.3-py3-none-any.whl
- Upload date:
- Size: 12.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via: uv/0.7.19
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
507d5c14e89773bbebc2d2994d9e8600e65ed954b3c4f6664ca038022b012c66
|
|
| MD5 |
b947c4ef4e9dfddf21d657a5591901ca
|
|
| BLAKE2b-256 |
96a01bb21f94996cd9b4dfd8a099d8ae5f56ca93c6579e5cd643dc29a2c0acc7
|