What problem does it solve?
Developers often need to call multiple Sentry tools from within an application without routing every action through a model or external planner. This Skill provides a single, typed Python client that uses MCP to invoke Sentry tools programmatically, reducing latency and token usage.
Core Features & Use Cases
- Typed async methods for each Sentry tool (e.g., whoami, find_organizations, find_projects, get_issue_details, search_events, and more) to simplify integration.
- OAuth-based authentication with optional custom providers and token persistence for reuse across sessions.
- Flexible parameter handling (organizationSlug, regionUrl, projectSlug, etc.) and robust result parsing (JSON when possible, raw text otherwise).
- Works as a self-contained client that can be integrated into Python apps, automation scripts, or agents requiring Sentry insights.
Quick Start
Create an instance of SentryApp and call a tool such as whoami to fetch the authenticated user.