AdStrike MCP
AdStrike is an Active Directory assessment MCP server that turns engagement material into HTML, Markdown and JSON reports.
- 455 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Dual-use tools can extract secrets and change domain state.
- An incorrect target can affect third-party systems outside the authorized scope.
Install
- Works in
Manual install
bash install.shRun as a normal user from an existing AdStrike checkout; installs dependencies and the local environment. MCP connection is a separate step.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Dual-use tools can extract secrets and change domain state.
- An incorrect target can affect third-party systems outside the authorized scope.
This is third-party code. Review the repository files before installing.
What it does
The AdStrike MCP server connects an agent to Active Directory security assessment tools. The host invokes domain assessment modules and receives results in a shared engagement session. That session retains target details, credentials and findings. The agent can use one result to select the next step. The project wraps external AD utilities and produces HTML, Markdown and JSON reports. Use is restricted to systems covered by written authorization.
Who it is for. Active Directory assessors with written authorization who need to combine results from external tools.
Good fit when
- A test-domain assessment collects findings in one session.
- An authorized engagement produces a report of findings.
- An analyst compares results from several utilities in a lab.
- A team uses an existing MCP client to manage authorized testing.
Not a fit when
- You lack written authorization for the target systems
- You only need guidance without executing tools
- You cannot share test-domain credentials with the agent
Example request
Prepare a plan for an authorized test-domain assessment. First list the scope and actions needing separate approval.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
This is a research beta: a module may fail on your target, so test it in an authorized lab before planning a live assessment. Kali Linux 2024 or newer, or Parrot OS, with Python 3.10 or later is recommended. The selected operations require their corresponding external tools to be installed. Credentials remain in the shared session and are reused by later calls, so account scope must match the authorized engagement.
How to disable. Disable AdStrike in the MCP client and stop its server process.
FAQ
Is an Anthropic API key required?
No. The MCP server uses the host’s model rather than running its own.
Where does installation run?
In an AdStrike checkout. The installer creates a repository-local virtual environment.
Should I run the installer as root?
No. A normal-user installation keeps project files owned by that user.
How do I connect the MCP server?
The AdStrike checkout includes a project `.mcp.json` with relative paths. For use from any directory, use absolute paths as described in the [MCP section](https://github.com/capture0x/AdStrike#mcp-server-no-api-key). Confirm that they point to your checkout before starting.
Does approving the MCP server authorize a test?
No. Server approval allows the connection; authorization to assess systems is separate.
Discussion
No questions or reviews yet.
Related
MCP serversAn offensive security MCP server: the agent runs 150+ pentest and recon tools and drives automated target assessments
Security scanner for MCP servers that looks for hardcoded keys, tool poisoning and injection flaws in repository code.
Security Detections MCP: a detection rule database
Security Detections MCP
An MCP server that merges Sigma, Splunk, Elastic and KQL rules into one database for finding detections and ATT&CK gaps
An MCP server for BloodHound: an agent analyzes Active Directory attack paths in plain language instead of hand-written Cypher queries to Neo4j
Security scanner for coding agents: checks code, MCP servers, prompts and model-invented packages, and gives an A to F grade.
MCP server for the ARL-Next platform: returns ports, fingerprints, certificates, vulnerabilities and leaks for a domain, IP or URL from your asset database
README in short
Commands, findings and output paths are stored with engagement context. Later calls in the same session reuse its saved credentials.