A purpose-built Model Context Protocol server that exposes a small, safe set of Splunk investigation tools to an AI agent. The agent reaches Splunk only through these tools, so every action it takes against the platform is named, typed and guard-railed at this boundary.
List the indexes available on the Splunk instance as JSON.
Run an existing saved search by name and return its results as JSON.
List saved searches (detections) defined on the instance as JSON.
Run a read-only SPL search and return the results as JSON. Use this to gather evidence: authentication records, process activity, network flows and DNS lookups. Destructive or data-moving commands are rejected. Returns JSON {spl, result_count, results}.
Check whether an SPL query is safe to run. Returns JSON {safe, reason}.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 0 | 2026-07-28+ | v2 |