Enterprise MCP gateway for governed, audited Cortex XSIAM access with tool-level policy enforcement, identity middleware, and audit logging
Server has 12 tools with READ_ONLY risk classification, but definition quality is significantly below production baseline. Most tool descriptions are present but lack actionable detail (avg ~80 chars, well below 194-char baseline). Parameter descriptions are sparse or missing entirely. Input schemas exist but lack proper type constraints and validation guidance. No output schemas documented. Error handling is absent from tool definitions. Tool naming follows verb_noun convention (positive), but parameter naming is inconsistent (e.g., 'filters' object lacks field documentation). No evidence of pagination support despite tools returning potentially large result sets. Security-critical tool (execute_xql_query) lacks proper guardrails in definition.
use only with the opaque cursor from a prior response
use when the requested concept has no matching discovered field. Arguments include dataset and optional field_name_contains
privileged raw-XQL escape hatch. Never use it for these standard-reader cases
Get security cases from Cortex XSIAM
Get guidance on dataset query construction and optimization
Get security issues from Cortex XSIAM
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. Violates pattern:tool and pattern:response-shaper.
Parameter descriptions are minimal or absent. 'filters' object in query_dataset and get_cases/get_issues lacks field documentation. LLMs cannot infer what filter fields are valid. Violates pattern:tool-description.
execute_xql_query marked as 'privileged raw-XQL escape hatch' but lacks permission gates or audit guidance in definition. No error handling documented. Violates pattern:permission-gate and pattern:audit-trail.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | <=2025-11-25 | v2 |
Get guidance on log search queries and best practices
Get help with XQL query syntax and functions
Get current XQL query quota and usage information
List available log datasets for querying
typed query for one explicit allowed dataset. Rows mode requires fields. Aggregate mode uses metrics, group_by, optional time_bucket, order_by, timeframe, and a small limit
Search logs using natural language or structured queries
No pagination support visible in tool definitions. get_cases and get_issues accept 'limit' but no offset/cursor or total_count guidance. Large result sets will blow context window. Violates pattern:paginated-result.
Tool descriptions lack actionable context. 'Get guidance on log search queries' (get_log_search_guidance) does not explain WHEN to call it or what it returns. Descriptions average ~80 chars vs 194-char baseline. Violates pattern:tool-description.