read-sentry-issuelisted

# Read Sentry Issue ## When to use - User shares a Sentry issue/event link `https://sentry.io/organizations/<org>/issues/<id>/` or `<org>.sentry.io/issues/<id>/`. - User mentions Sentry or error tracking by name. - Investigating an exception, stack trace, or production error that might be tracked in Sentry. ## Authentication Use Sentry's official hosted MCP server. It needs no manually-issued token: the first call opens a browser to sign in, and Sentry's OAuth session is reused for later calls. - Discover the available tools with `ToolSearch` (query `"sentry"` or a specific tool name like `"get_issue_details"`), exact registered names depend on how the server was added locally, do not hardcode a guess. - One-time setup, if the tools aren't found: tell the user to run `claude mcp add --transport http sentry "https://mcp.sentry.dev/mcp?skills=inspect" -s user`, then `claude mcp login sentry` (same URL works for OpenCode or any other MCP-capable client, this is Sentry's own public endpoint, not org-specific). - The `?skills=inspect` URL param alone is NOT enough, verified live: Sentry's MCP always exposes a small fixed "core" tool set (`find_organizations`, `find_projects`, `search_issues`, `search_events`, `get_sentry_resource`, `search_sentry_tools`, `execute_sentry_tool`, plus `update_issue` and `analyze_issue_with_seer`) regardless of that param, it only narrows the secondary catalog reached through `search_sentry_tools`. The real read/write boundary is decided on the O