Skip to content

README.md

Latest commit

 

History

History
61 lines (37 loc) · 2.67 KB

README.md

File metadata and controls

61 lines (37 loc) · 2.67 KB
Raw

TierZero for Cursor

Agentic production engineering inside Cursor — for SWE, SRE, and DevOps work. The plugin registers TierZero as an MCP server so the agent can resolve and investigate issues against your existing observability stack, CI/CD systems, and self-improving knowledge bases.

What you get

  • MCP server — tierzero is registered automatically. The agent can call every tool the TierZero MCP server exposes (tierzero_ask, tierzero_fetch_context, plus any others the server adds in the future) without shelling out.
  • Slash command — /investigate <query> to kick off a TierZero investigation.
  • Skills — tierzero-investigate (the entry point for any production problem — alerts, errors, latency, deploy regressions) and tierzero-fetch (load a saved TierZero chat, investigation, or artifact). The agent picks them automatically based on context.
  • Rule — alwaysApply rule that nudges the agent to reach for TierZero whenever real production behavior would change its answer.

Requirements

A TierZero account with at least one integration connected (Datadog, New Relic, CloudWatch, Grafana Cloud, Sentry, GitHub, BuildKite, Temporal, Confluence, Notion, Slack, etc.).

Authentication

The plugin uses OAuth — no tokens, no env vars. On first use Cursor opens a browser to TierZero, you log in, consent to the requested scope, and the token persists in Cursor. To revoke access, remove the MCP server from Cursor Settings → MCP.

Usage

Run an investigation

/investigate what errors hit payment-service in the last hour?

The agent calls tierzero_ask and summarizes the result.

Load saved context

Just paste a TierZero URL or artifact UUID into the chat:

https://app.tierzero.ai/investigations/SW52ZXN0...
bf904904-afdc-4cf2-94d8-76a4a8bb4f75

The tierzero-fetch skill triggers automatically — the agent loads the chat / investigation / artifact and folds it into the current task.

Hands-off

You don't have to invoke explicitly. The included rule tells the agent to reach for TierZero whenever production data would change its answer, e.g. while debugging an incident, reviewing a hot-path PR, or designing a feature that depends on current traffic patterns.

What can the agent see?

Everything your TierZero org has connected. Common integrations:

  • Logs / metrics / traces — Datadog, New Relic, CloudWatch, Grafana Cloud, Sentry
  • Code & CI/CD — GitHub, BuildKite, GitHub Actions
  • Workflows — Temporal
  • Knowledge — Confluence, Notion, Slack

Documentation

License

MIT