monoscope-skills

Claude Code skills for the Monoscope observability platform. Gives Claude the knowledge to use the monoscope CLI to investigate production incidents, triage issues and alerts, write KQL queries, and manage your observability resources.

Installation

Claude Code

claude plugin marketplace add monoscope-tech/skills
claude plugin install monoscope-skills@monoscope-skills

Restart Claude Code after installation. Skills activate automatically when relevant.

Update:

claude plugin marketplace update
claude plugin update monoscope-skills@monoscope-skills

Or run /plugin to open the plugin manager.

Skills Package (skills.sh)

For agents supporting the skills.sh ecosystem:

npx skills add monoscope-tech/skills

Works with Claude Code, Cursor, Cline, GitHub Copilot, and other compatible agents.

Prerequisites

Install the monoscope CLI and authenticate:

curl monoscope.tech/install.sh | sh
monoscope auth login

Set your default project (or pass MONOSCOPE_PROJECT per-session):

monoscope config set project <your-project-uuid>

Available Skills

Skill	Description
kql-reference	Full KQL query language reference
investigate	Investigate production incidents
triage	On-call sweep: issues, alerts, log patterns
manage	Monitors & dashboards as code (create alerts, GitOps apply)
instrument	Instrument an app with OpenTelemetry auto-instrumentation

kql-reference

Complete reference for Monoscope's KQL dialect — operators (==, !=, has, contains, startswith, matches, in, has_any …), aggregations (count, dcount, percentile …), time binning (bin_auto, bin), scalar functions, the full telemetry schema, and 15+ example queries.

Consult this whenever writing, reading, or debugging a KQL query.

investigate

Investigate production incidents using the monoscope CLI. Walks Claude through discovering active services, searching logs and traces, inspecting trace trees, watching live events, and checking metrics against SLO thresholds.

Example prompts:

• "Investigate the 500 errors in the payment service over the last hour"
• "Find logs related to the timeout we saw at 10:34 UTC"
• "Check the error rate for checkout-api and show me what's failing"
• "What happened to the auth service between 14:00 and 14:30?"

triage

On-call sweep: review open issues, acknowledge noisy log patterns, mute or resolve flapping monitors, and bulk-action alert queues.

Example prompts:

• "Triage all open issues for the payment service"
• "Do an on-call sweep and clear the alert queue"
• "Mute the noisy database connection monitor for 30 minutes"
• "Bulk acknowledge the runtime exception issues from last night"

manage

Create and manage alert monitors and dashboards as code: YAML files upserted with monoscope monitors apply / dashboards apply (idempotent — keyed by title / file_path), plus the yaml dump → edit → apply round-trip.

Example prompts:

• "Alert me when checkout errors exceed 10 in 15 minutes"
• "Create a latency dashboard for the payment service"
• "Put our monitors in the repo and apply them from CI"
• "Add a p99 widget to the on-call dashboard"

instrument

Add OpenTelemetry instrumentation to an application so traces, logs, and metrics flow into Monoscope. Detects the language and framework, installs the right Monoscope SDK or OTel packages, writes the SDK init code and middleware wiring, sets the OTLP exporter env vars, and verifies data is flowing with monoscope services list.

Supports native Monoscope SDKs (with request/response capture, error reporting, and outgoing-request monitoring) for: Express, Fastify, NestJS, FastAPI, Django, Gin, Echo, Chi, Fiber, Spring Boot, Laravel. Falls back to standard OTel auto-instrumentation for Ruby, Rust, .NET, Flask, Pyramid, and any other OTel-supported language.

Example prompts:

• "Instrument this Express app with Monoscope"
• "Add OpenTelemetry tracing to my FastAPI service"
• "Wire up observability for my Go Gin API"
• "Set up the OTLP exporter for my Spring Boot app"
• "Get my Django app sending data to Monoscope"

Environment variables

Variable	Description
MONOSCOPE_API_KEY	API key — takes precedence over stored token
MONOSCOPE_PROJECT	Default project UUID
MONOSCOPE_API_URL	API base URL (default: https://api.monoscope.tech). Point at http://localhost:8080 to drive a self-hosted dev server from these skills.
MONOSCOPE_DEBUG	Set to 1 to print every outgoing request URL+params to stderr — invaluable when an agent gets a 4xx and needs to inspect what it actually sent. Same as the global --debug flag.
MONOSCOPE_FORCE_COLOR	Set to 1 to keep TTY-style (table) output even when stdout is piped.

Output format

Output format is auto-detected from stdout: TTY → table, pipe/redirect → JSON — so agents piping output get JSON automatically. Force a format with the global --json, --yaml, or --table flags (precedence --json > --yaml > --table). -p/--project PID overrides the project per-invocation.

Output shapes (stable across versions)

The skills assume these JSON envelopes — they're pinned by integration tests in test/integration/CLI/CLIE2ESpec.hs so a regression breaks CI rather than the skill in production.

Command	Shape
events search, logs search, traces search, events get, events context	{events: [{id, timestamp, service, summary, trace_id, kind, ...}], count, has_more, cursor}
services list	{services: [{name, events}], count} (precomputed via the server's facet store — fast even on huge projects)
<resource> list (issues, monitors, dashboards, api-keys, teams, members, endpoints, log-patterns)	{data: [...], pagination: {has_more, total, cursor, page, per_page}} — use .data[] in jq
auth status (piped or --json)	{authenticated, method, api_url, project}
facets [FIELD]	{<field_path>: [{value, count}, ...]}
events context --summary	base envelope plus traces: [{trace_id, services, span_count, error_count}, ...]

Links

• Monoscope — observability platform
• CLI reference — full monoscope command reference
• monoscope-tech/monoscope — main repo

License

Apache-2.0