Notifications
Clear all
Topic starter
21/08/2026 4:45 pm
TL;DR: Anthropic’s directory listing makes a security graph easier to query from Claude, so teams can ask natural-language questions about exposure, toxic combinations, and blast radius without rebuilding their stack, according to JupiterOne. The governance issue is not convenience but whether graph-backed context becomes an AI workflow input without clear identity and access boundaries.
NHIMG editorial — based on content published by JupiterOne: JupiterOne Joins Anthropic's MCP Directory: Bring Your Security Graph Into Every AI Workflow
Questions worth separating out
Q: How should security teams govern AI access to a security graph?
A: Treat AI access to a security graph like any other privileged integration.
Q: Why does natural-language querying create governance risk for IAM teams?
A: Natural-language querying lowers the barrier to broad data access, which can expand who asks what and how often.
Q: What should teams check before connecting AI tools to operational security data?
A: Check whether the AI client has explicit scope limits, whether the data returned is redacted where necessary, and whether query activity is auditable.
Practitioner guidance
- Govern MCP access as a privileged integration Define which AI clients can reach the security graph, what data domains they can query, and which response fields must be filtered or suppressed.
- Apply least-privilege to graph relationships Limit assistants to the minimum asset, identity, and vulnerability relationships required for the workflow, rather than exposing the entire graph by default.
- Log AI-mediated security queries end to end Record the prompt, the queried objects, the returned context, and the downstream action so review teams can trace how sensitive security data was used.
What's in the full article
JupiterOne's full blog post covers the operational detail this post intentionally leaves for the source:
- How to connect Claude to the JupiterOne MCP server and authenticate with an existing account
- The specific questions the security graph can answer in live triage workflows
- Examples of the graph relationships the integration can return for exposure and blast-radius analysis
- The implementation path for teams that want to use the existing JupiterOne instance without rebuilding their stack
👉 Read JupiterOne's blog post on Anthropic's MCP directory listing for JupiterOne →
JupiterOne in Anthropic's MCP directory: what changes for IAM teams?
Explore further
21/08/2026 5:05 pm
MCP is becoming a control-plane issue, not just an integration pattern. When a security graph is exposed through an AI-readable protocol, the identity question is no longer only who can log in. It is what the AI client can query, infer, and reuse at runtime. That moves MCP from a developer convenience into identity governance territory, because the access boundary now spans tools, context, and downstream interpretation.
A few things that frame the scale:
- Only 1.5 out of 10 organisations are highly confident in their ability to secure NHIs, compared to nearly 1 in 4 for securing human identities, according to The State of Non-Human Identity Security.
- The average estimated time to remediate a leaked secret is 27 days, according to The State of Secrets in AppSec.
A question worth separating out:
Q: How do security graphs change blast-radius analysis for practitioners?
A: They move blast-radius analysis from isolated findings to relationship-driven exposure. Instead of asking only whether an asset is vulnerable, teams can ask what the asset touches, what identity can reach it, and what the next-hop impact would be. That improves triage, but only if the graph access path is tightly governed.
👉 Read our full editorial: JupiterOne MCP directory listing expands security graph access
