Observability Skills for OpenClaw Agents

agent-bom discover snowflake

@msaad00

Discover Snowflake Cortex, Snowpark, notebook, Streamlit, MCP, and AI-observability assets from the operator's environment, emit canonical agent-bom inventory JSON, and scan it without giving agent-bom long-lived Snowflake credentials. Use when a user asks to inventory Snowflake AI or Cortex infrastructure as canonical inventory.

DebugBundle

@debugbundle

Use DebugBundle for runtime error reporting, crash reporting, incident reporting, incident response, live app monitoring, and production monitoring focused on runtime failures, customer-facing incidents, and endpoint health—not generic infrastructure metrics. Investigate exceptions, alerts, logs, observability signals, health checks, and debug bundles; review product analytics; and guide evidence-based fixes through CLI primarily, with MCP where needed.

Grafana

@byungkyu

Grafana API integration with managed authentication. This is a write-capable integration — it can read, create, update, and delete dashboards, data sources, folders, annotations, and teams in your Grafana instance, read alert rules and org info, read the current user's own profile, list installed plugins, and search service accounts (read-only inventory of identities that hold API tokens - no service account or token is created or changed here). Alert rules are documented read-only; the provisioning namespace can also write them, and this skill does so only on a specific user request with confirmation. Use this skill when users want to interact with Grafana for monitoring, visualization, and observability. All write operations (creating/updating/deleting dashboards, folders, data sources, or teams) require explicit user approval with specific resource identifiers before execution. Team administration is included: creating, updating, and deleting teams changes who can reach which dashboards and folders, so treat those as identity- and access-affecting writes that need explicit confirmation of the team by id and name. It does not cover user accounts, organization membership, licensing, or server-level platform settings. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway). Calls run through the `maton` CLI with OAuth login, or over raw HTTP with a Maton API key where the CLI cannot be installed. Every call is authenticated as the user's connection and reaches only what that connection's authorization allows, which the provider enforces on every request; the endpoints documented here are the ones this skill uses, and any other endpoint of this app needs the user to ask for it by name. Default to read and list calls, and confirm every write or new connection with the user. This file also documents the three constructs that turn a Grafana connection into automation, in the order they are used: the connection (the first step), a hosted function that runs a Grafana action through the Maton SDK, and a trigger that calls that function on a schedule or on an event. Those sections are the platform's own reference text, shared with the api-gateway skill, with Grafana examples; they add no Grafana capability - Grafana is not an event source, a trigger cannot read Grafana data, and the files under `references/<source>/triggers.md` are the platform's event catalogues for the sources Maton offers (time, Calendly, GitHub, Gmail, HubSpot, Linear, Notion, Slack, Stripe).

New Relic CLI Skills

@vince-winkintel

Monitor, query, and manage New Relic observability data via the newrelic CLI. Covers NRQL queries, AI Monitoring telemetry, APM performance triage, deployment markers, read-only alert and incident inspection, infrastructure monitoring, agent diagnostics, and identity/access administration. Use when user asks about LLM token usage or AI application telemetry, application performance, error rates, slow transactions, deployment tracking, incident inspection, New Relic setup, users, groups, roles, permissions, or access grants.

kibana

@byungkyu

Kibana API integration with managed authentication. This is a write-capable integration — it can read, create, update, and delete saved objects, dashboards, data views, spaces, alerts, fleet resources, connectors/actions, security roles, and cases in your Kibana instance. All write operations require explicit user approval with specific resource identifiers before execution. Connector execution triggers external side effects (e.g., sending emails, invoking webhooks) — always confirm the connector ID, target, and payload with the user before executing. Use this skill when users want to interact with Kibana for observability, security, and search analytics. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway). Calls run through the `maton` CLI with OAuth login, or over raw HTTP with a Maton API key where the CLI cannot be installed. Every call is authenticated as the user's connection and reaches only what that connection's authorization allows, which the provider enforces on every request; the endpoints documented here are the ones this skill uses, and any other endpoint of this app needs the user to ask for it by name. Default to read and list calls, and confirm every write or new connection with the user. This file also documents the three constructs that turn a Kibana connection into automation, in the order they are used: the connection (the first step), a hosted function that runs a Kibana action through the Maton SDK, and a trigger that calls that function on a schedule or on an event. Those sections are the platform's own reference text, shared with the api-gateway skill, with Kibana examples; they add no Kibana capability - Kibana is not an event source, a trigger cannot read Kibana data, and the files under `references/<source>/triggers.md` are the platform's event catalogues for the sources Maton offers (time, Calendly, GitHub, Gmail, HubSpot, Linear, Notion, Slack, Stripe).

alibabacloud-cms-manage

@sdk-team

Entry skill for the aliyun CLI distribution of CloudMonitor (CMS). Use when the user mentions aliyun cms2, CloudMonitor, CMS commands, or any CMS module operation such as Integration Policy/Center, APM, RUM, Prometheus Service, Recording rule, alert rule, alert template, alert history, event hub, SLS event, PromQL, cloud resource, service observability, monitoring onboarding, metric query, etc.

OpenClaw Usage Report

@dtsola

OpenClaw usage and performance reporting. Parse session JSONL to answer how long each agent task took, which tools/skills/models were used, and how many tokens were consumed — zero dependency, local only, no cost dimension (the script reads, computes and outputs no cost fields at all; token is the primary metric). Read-only: never modifies any file. Outputs are aggregated statistics only: no conversation content, no session identifiers (session-level detail is opt-in via --include-sessions) and no credential fields. The time window is a hard boundary: with a window selected (or the default "today") every dimension aggregates in-window events only, and full history requires the explicit --all flag. Documentation, report and JSON output follow the same scope; links in the docs are human-readable references, the runtime script makes no network calls. Report text defaults to Chinese and can be adapted to the user's language; the skill imposes no language or region restriction. Use when the user asks about token usage, task duration, slowest tools, skill usage, or per-agent consumption (今天花了多少 token/哪个工具最慢/ 任务耗时/用量报告), or scheduled via cron. 中文：OpenClaw 用量与性能查询。 解析 session JSONL，回答每次 agent 任务耗时、所用工具/技能/模型、token 消耗。零依赖纯本地，不提供成本维度（token 为主指标）。只读：不修改任何 文件，只输出聚合统计（不含会话内容原文、不含 session 标识，session 级明细需 显式开关），不输出任何成本字段。用户问 token 用量、任务耗时、最慢工具、 技能使用、按 agent 消耗时使用。cron 每日日报为可选项，须用户显式要求 并确认后自行设置。

Kubernetes Agent Swarm

@kcns008

Kubernetes & OpenShift Platform Agent Swarm — A coordinated multi-agent system for cluster operations. Includes Orchestrator (Jarvis), Cluster Ops (Atlas), G...

golang-observability

@samber

Golang everyday observability — the always-on signals in production. Covers structured logging with slog, Prometheus metrics, OpenTelemetry distributed tracing, continuous profiling with pprof/Pyroscope, server-side RUM event tracking, alerting, and Grafana dashboards. Apply when instrumenting Go services for production monitoring, setting up metrics or alerting, adding OpenTelemetry tracing, correlating logs with traces, migrating legacy loggers (zap/logrus/zerolog) to slog, adding observability to new features, or implementing GDPR/CCPA-compliant tracking with Customer Data Platforms (CDP). Not for temporary deep-dive performance investigation (→ See `samber/cc-skills-golang@golang-benchmark` and `samber/cc-skills-golang@golang-performance` skills).

Huawei Cloud Cce Kubernetes Event Analyzer

@pintudeyudi

Query and analyze Kubernetes events in Huawei Cloud CCE clusters using API or LTS logs, apply filters, identify patterns, and hand off to diagnosis skills.

aws-agentcore-langgraph

@killerapp

Deploy production LangGraph agents on AWS Bedrock AgentCore. Use for (1) multi-agent systems with orchestrator and specialist agent patterns, (2) building stateful agents with persistent cross-session memory, (3) connecting external tools via AgentCore Gateway (MCP, Lambda, APIs), (4) managing shared context across distributed agents, or (5) deploying complex agent ecosystems via CLI with production observability and scaling.

grafana-foundation-sdk

@tenequm

Build Grafana dashboards as code with the grafana-foundation-sdk typed builders (TypeScript or Go). Use when creating, modifying, or generating Grafana dashboard JSON programmatically, converting hand-written dashboard JSON to typed code, building monitoring dashboards, or working with Prometheus/Loki queries in dashboards.

Huawei Cloud Cce Alarm Correlation Engine

@pintudeyudi

Huawei Cloud AOM alarm correlation analysis skill for CCE operations. Use this skill when the user wants to: (1) query AOM active and historical alarms, (2)...

claw-observability

@rdmss

Real-time visual observability for Claude Code multi-agent workflows. Automatic reporting via hooks — zero agent involvement.

Huawei Cloud Cce Metric Analyzer

@pintudeyudi

Huawei Cloud CCE Metric analysis skill using Python SDK dispatcher. Use this skill when the user wants to: (1) query Pod/Node CPU/memory/disk metrics, (2) ge...

Grafana

@hith3sh

Grafana HTTP API integration. Inspect dashboards, folders, data sources, alert rules, and teams. Use this skill when users want to query observability data,...

New Relic

@hith3sh

Monitor applications, infrastructure, logs, synthetic checks, and cloud services in New Relic. Configure alerts, manage dashboards, query telemetry data with...

Kibana

@hith3sh

Manage Elastic Kibana for observability and security operations. Query data views, manage alerting rules and detection engine rules, handle Fleet agent polic...

Datadog

@hith3sh

Inspect Datadog monitors, metrics, logs, incidents, dashboards, and observability data - powered by ClawLink.

agentic-ledger

@shekharbhardwaj

Track every token and dollar your OpenClaw agent spends, set budget walls that actually refuse calls, and replay runs on free local models. Local-first: all data stays on this machine. Use when the user asks about cost, spend, tokens, budgets, usage reports, or what the agent has been doing.

RelayPlane

@relayplane

Agent ops layer for OpenClaw — observability, governance, and cost optimization with automatic failover. Never breaks your setup.

LLMxRay: diagnose your Ollama calls

@lognebudo

Find out what went wrong with your own Ollama calls and fix it, using LLMxRay's local, read-only MCP tools. Use when a local model is slow, a prompt is rejected or silently cut for exceeding the context, the prompt cache seems lost between turns, a model runs partly on CPU, tool schemas crowd the prompt, or you must choose which local model and context length fit in video memory. Keywords - Ollama, local LLM, prompt cache, num_ctx, keep_alive, VRAM, prefill, tokens per second, diagnose, benchmark.

box-integration-best-practices

@iskysun96

Build, debug, or review repeatable Box API integrations and agent workflows for efficient calls, bulk work, and 429 handling.

cron-audit

@cprite

Audit OpenClaw cron jobs for failures that stay silent — announce delivery that cannot resolve a recipient (channel "last" in an isolated session, Discord with no "channel:" target), error streaks nobody saw, output or failure alerts that were never delivered, bestEffort muting, jobs without failure alerts, overdue schedules and jobs that never ran. Use when the user asks why a cron job "does nothing", whether scheduled jobs are healthy, wants a cron health check or watchdog, or before re-enabling an old job.

agent-ops-monitoring

@deciqai

Activate when: user says 'how do we watch the agent after launch', 'AI output quality is getting worse', 'inference spend suddenly jumped', 'the agent broke something, how do we roll back', 'we need an AI ops weekly review', or an AI agent is already in production handling real business. Do NOT activate when: the project is still at the business-case stage (use ai-roi-business-case); usage is one-off and manual with no ongoing traffic. More: deciqai.com/c/agent-ops-monitoring