REPO PACKAGE
signoz agent-skills
COMMUNITY12skills
2ready to use
LABSCO SUMMARY
Alerts and dashboards each get three: create, explain, and either investigate (alerts) or modify (dashboards). Two skills cover queries — signoz-generating-queries for ad-hoc exploration, and signoz-writing-clickhouse-queries for raw SQL a dashboard panel's builder UI can't express. One skill manages saved Explorer views, one searches SigNoz's own documentation, and one — signoz-setting-up-observability — is an orchestrator that sequences the single-purpose skills into a full SLI/SLO-aware setup once telemetry is already flowing. The twelfth, signoz-mcp-setup, is what the README tells you to run first whenever the others' signoz_* tools stop responding.
This is for a team that already has telemetry flowing into SigNoz Cloud or a self-hosted instance and wants an agent that creates the alert or dashboard directly rather than walking through the UI — it does no instrumentation itself. Every one of the twelve depends on the SigNoz MCP server actually being connected: nine are flagged in our records as needing that live connection, and the setup skill itself needs a Cloud region or a self-hosted URL and account before anything else works.
READ THE FULL ANALYSIS
The plugin changes your permissions on install. The Claude Code version of this package ships a PreToolUse hook that auto-allows WebFetch calls to signoz.io domains — the README says so plainly, but it is still a real, automatic change to what your assistant is allowed to do, and it does not cover Bash-based fetches like curl or wget, which still ask for permission normally.
Checked 19 September 2026 from the repository README and the 12 skills currently live in our database; the package is MIT-licensed.
ALSO IN THIS PACKAGE
PLUGINS
1 itemPLUGIN
signoz
Official SigNoz plugin for MCP setup, docs, queries, dashboards, and alerts
WHAT'S INSIDE
12 showing · 12 totalReady to use
Nothing else to set up — install it and go.
12 of 12
SETUP
Showing all 12 skills
signoz-creating-alerts
Builds a new SigNoz alert from a plain description of what to watch, checking that the target data actually exists before saving so the alert doesn't end up silently never firing.
signoz-creating-dashboards
Watching a database, a Kubernetes cluster or an AI app normally means assembling a screen of charts by hand; here a ready-made one is imported when a matching template exists, and built panel by panel when none does.
signoz-explaining-alerts
Reads an existing SigNoz alert's configuration and translates it into plain language: what it watches, when it fires, where it notifies, and whether it has actually been firing.
signoz-explaining-dashboards
Someone who has been handed a monitoring screen they did not build gets a walkthrough of it: what each chart measures, what the dropdowns change, and what a healthy or unhealthy reading looks like.
signoz-generating-queries
Asks the monitoring data a one-off question — how slow, how many, which service is worst right now — and simply answers it, with nothing left saved behind.
signoz-investigating-alerts
An alert has gone off and nobody knows why; this pulls together everything else that was happening at that moment — other measurements, individual requests, log lines — and comes back with a ranked list of likely causes and the evidence for each.
signoz-managing-views
Turns a SigNoz Explorer search into a saved, named view the whole team can reopen later, and can also list, rename, update, or delete views that already exist.
signoz-mcp-setup
Connects or repairs the SigNoz MCP server registration for whichever coding tool you're using, so its SigNoz query, dashboard, alert, and view tools actually work.
signoz-modifying-dashboards
Changes an existing SigNoz dashboard — adding, removing, or editing panels, variables, or layout, or renaming it — by fetching its current state first and validating every changed query before saving.
signoz-searching-docs
Looks up an answer in SigNoz's official documentation and cites the exact page it came from, instead of guessing from memory.
signoz-setting-up-observability
Once a service is already sending its data to SigNoz, this does the rest of the setup in one pass: the targets it should be held to, the alerts for missing them, the screens to watch it on, and the searches worth keeping.
signoz-writing-clickhouse-queries
Writes the raw ClickHouse SQL for a SigNoz dashboard panel when the query is too complex for the drag-and-drop builder, such as a custom join, a window function, or a regex pulled out of a log body.
No skill in this package matches that.
HOW TO GET IT
WHOLE PACKAGEInstall the whole package (12 skills)
npx skills add signoz/agent-skillsSINGLE SKILLOr install a single skill
npx skills add signoz/agent-skills --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.
PLUGIN MARKETPLACEAdd the marketplace, then install the plugin
/plugin marketplace add signoz/agent-skills/plugin install signoz@signoz-skillsTyped inside the agent's own prompt, not in a terminal. The marketplace is called signoz-skills, which is the part after the @.