Agent Skills & Plugin

SigNoz Cloud - This page applies to SigNoz Cloud editions.
Self-Host - This page applies to self-hosted SigNoz editions.

Agent Skills are an open format for extending AI coding assistants with specialized knowledge and capabilities. Once installed, your agent uses them whenever it detects a relevant task - no prompting required.

When you ask your AI assistant something like "why did this alert fire," "create a dashboard for my Postgres database," or "write a ClickHouse query for these logs," the agent picks the matching skill and follows its instructions.

SigNoz ships these skills two ways:

  • The SigNoz plugin - the recommended path. One install bundles every skill plus the MCP server registration, so your agent can both reason about SigNoz and act on your data. Available for Claude Code, Codex, Cursor, and Gemini CLI.
  • Individual skills via skills.sh - install one or more skill files on their own, for any skills.sh-compatible agent.

Install the plugin

Add the marketplace and install the plugin:

/plugin marketplace add SigNoz/agent-skills
/plugin install signoz@signoz-skills

On install, Claude Code prompts for your SigNoz Cloud Region (defaults to us; one of us, us2, eu, eu2, in, in2). Find your region under Settings → Ingestion in SigNoz, or see the region reference.

Then run /mcp, select the signoz server, and complete authentication.

To update after new releases:

/plugin marketplace update
/plugin update signoz@signoz-skills

Install individual skills

If you want only specific skills, or you use an agent without plugin support, install skill files directly with skills.sh. This works for any compatible agent (Codex, Cursor, Gemini, OpenCode, and more).

Install all SigNoz skills:

npx skills add SigNoz/agent-skills

Install a specific skill:

npx skills add SigNoz/agent-skills --skill signoz-searching-docs
npx skills add SigNoz/agent-skills --skill signoz-writing-clickhouse-queries

skills.sh installs the skill files only, not the bundled MCP registration. Skills that act on your data (queries, dashboards, alerts, views) need the SigNoz MCP Server connected separately.

Available Skills

The plugin ships the skills below. Your agent selects the right one automatically when it detects a matching task. Click any name to read its source SKILL.md.

Setup and docs

  • signoz-mcp-setup - Initialize or repair the SigNoz MCP server configuration for Claude Code, Codex, Cursor, VS Code/GitHub Copilot, Claude Desktop, Gemini CLI, Windsurf, Zed, Antigravity, OpenCode, or another MCP client.
  • signoz-searching-docs - SigNoz docs guidance for instrumentation, setup, querying, alerts, and APIs.

Querying your data

Dashboards

  • signoz-creating-dashboards - Create a new SigNoz dashboard from a natural-language intent - import a curated template (PostgreSQL, Redis, JVM, k8s, APM, LLM, etc.) when one fits, or build a custom dashboard with metric, trace, and log panels.
  • signoz-modifying-dashboards - Modify an existing SigNoz dashboard: add, remove, or edit panels, variables, queries, and layout.
  • signoz-explaining-dashboards - Explain panels, queries, and layout of an existing SigNoz dashboard.

Alerts

  • signoz-creating-alerts - Create SigNoz alert rules for threshold breaches, error rates, latency, anomaly detection, and absent-data conditions across metrics, logs, and traces.
  • signoz-explaining-alerts - Explain and interpret an existing SigNoz alert rule's configuration, evaluation behavior, notification routing, and recent fire frequency.
  • signoz-investigating-alerts - Diagnose why a SigNoz alert fired by correlating its signal with neighbor metrics, traces, and logs around the fire window, and ranking likely causes.

Views and workflows

  • signoz-managing-views - Create, list, inspect, update, or delete SigNoz saved Explorer views (logs, traces, metrics) via the SigNoz MCP server.
  • signoz-setting-up-observability - Orchestrate the full post-ingestion observability setup for a service - SLI/SLO capture, RED/USE exploration, focused dashboards, saved views, burn-rate and absent-data alerts, and a tuning loop - sequencing the single-artifact skills into one SLO-aware workflow.

Learn more

Last updated: June 14, 2026

Edit on GitHub

Was this page helpful?

Your response helps us improve this page.