Security 🧪 Community ✓ Claude Code ✓ Claude Desktop Standalone • SaaS Alerts
Saas Alerts
SaaS Alerts - SaaS security monitoring and alerting for M365 / Google Workspace: alerts, events, anomaly detection, multi-tenant response
Installation
Install this plugin individually:
/plugin marketplace add wyre-technology/msp-claude-plugins --plugin saas-alerts Or install all MSP plugins at once:
/plugin marketplace add wyre-technology/msp-claude-plugins Features
- Triage
Skills
This plugin provides 2 skills that teach Claude about Saas Alerts:
| Skill | Description |
|---|---|
triage | Triaging the SaaS Alerts queue across managed M365 / Google Workspace tenants: the triage tool surface, the critical-first sweep, per-customer summary and cross-tenant pattern workflows, the low/medium/critical severity model and its default dispositions, and the edge cases — legitimately empty results, time-window sensitivity, whitelist suppression, and per-partner rate limits. |
api-patterns | SaaS Alerts MCP fundamentals: API-key authentication via the gateway header, the MSP → customer → account → user hierarchy, navigation and functional tool naming, event filter parameters, cursor pagination, and HTTP error codes. |
Agents
This plugin provides 1 agent for autonomous task execution:
| Agent | Description |
|---|---|
saas-alerts-analyst | Use this agent when investigating and triaging SaaS Alerts security alerts across managed M365 / Google Workspace tenants — reconstructing what fired, attributing it to a user/tenant, judging severity, and recommending response. |
Commands
Available slash commands:
| Command | Description |
|---|
API Reference
| Base URL | |
| Authentication | |
| Rate Limit | |
| Documentation |
Example Usage
Using Skills
/skill saas-alerts:triage
Triaging the SaaS Alerts queue across managed M365 / Google Workspace tenants: the triage tool surface, the critical-first sweep, per-customer summary and cross-tenant pattern workflows, the low/medium/critical severity model and its default dispositions, and the edge cases — legitimately empty results, time-window sensitivity, whitelist suppression, and per-partner rate limits.