GuanceCloud

Generate, repair, or review Guance Dashboard JSON from real metrics CSV files, tag metadata, and resource-catalog or custom-object CSV/JSON data.

Generate Guance monitor JSON from CSV metric files for any component.

Convert Grafana dashboard JSON into Guance dashboard JSON with a fully self-contained skill package. Use when the user wants to analyze Grafana dashboards before conversion, run conversion, audit conversion gaps, improve units or PromQL compatibility with LLM-assisted review, repair panel or variable mappings, preserve settings, groups, and vars as much as possible, validate generated Guance dashboard JSON against the bundled schemas, or debug why a Grafana dashboard does not convert cleanly.

Use the owl CLI for Guance queries, diagnostics, root-cause analysis, and structured Markdown reports.

Convert Prometheus alerting rules into Guance monitor JSON and validate all generated DQL.

Generate, fix, explain, and review DQL; final executable DQL must pass dqlcheck item by item.

Convert, validate, and explain Alibaba Cloud SLS queries as GuanceDB DQL.

Use Trivy to perform authorized, scan-only security assessment of Kubernetes or service clusters, including all namespaces/workloads, all language application dependencies discoverable in images or source paths, container images, image configuration, Kubernetes resource configuration, secrets, misconfigurations, vulnerabilities, compliance posture, runtime/node collector findings, automatic ephemeral portable Trivy download with checksum verification and post-scan cleanup, and latest official vendor remediation guidance with citations. Trigger on Trivy, vulnerability scan, cluster scan, Kubernetes scan, image scan, container config scan, runtime scan, official vulnerability remediation, Chinese requests for vulnerability scanning, cluster scanning, image scanning, container configuration scanning, runtime scanning, and official remediation guidance. Forbid remediation, patching, apply/delete/upgrade actions, exploit testing, or configuration changes unless the user explicitly confirms a separate fix request.

Create, review, and validate GuanceCloud Arbiter scripts for security detection rules. Use when Codex needs to author SIEM or CSPM detection logic, query data with dql or PromQL, process DQL series, call trigger to generate security events, inspect Arbiter built-in functions, or debug Arbiter syntax/runtime/trigger-output problems.

生成、修复、解释、评审 DQL。凡是要交付“最终可执行 DQL”，都必须先逐条通过 dqlcheck 校验，再输出最终结果。

Create, refine, and validate GuanceCloud/DataKit pipeline scripts from raw message samples. Use when Codex needs to author pipeline .p scripts, choose json/load_json versus grok extraction for text or JSON messages, run pipeline-go validation, or debug parsing/runtime/extraction errors.

Write a new library instrumentation end-to-end. Use when the user ask to add a new APM integration or a library instrumentation.

Converts Spock/Groovy test files in a Gradle module to equivalent JUnit 5 Java tests. Use when asked to "migrate groovy", "convert groovy to java", "g2j", or when a module has .groovy test files that need to be replaced with .java equivalents.

Use when debugging with an external DataKit bug-report bundle such as info-<timestamp> or info-<timestamp>.zip, and you need to use the current bundle as the primary evidence while treating docs/bug-report as historical hints

Use when the user wants to install or upgrade the OpenClaw OTEL plugin through the GTrace install flow, and expects the agent to build and run the install command from explicit parameters such as installer URL, endpoint, token, version, and tags. This skill is only for the GTrace install path.