数据库实例巡检与报告生成

数据库实例巡检与报告生成，支持配置检查、性能检查、安全检查、报告生成、智能巡检、异常检测、根因分析、风险预测。 使用场景： - 用户说"巡检" → 执行 run - 用户说"生成报告" → 执行 report - 用户说"检查配置" → 执行 run --type configuration - 用户说"建立基线...

Security Scan

The skill's instructions assume a local CLI (dbskiter) and access to database credentials/metrics, but the package declares no required binaries, install steps, or credential/environment requirements — that's an internal mismatch you should clarify before using it.

Purpose & Capability

The SKILL.md consistently instructs the agent to run a local CLI named `dbskiter` (many example commands). Yet the skill metadata declares no required binaries and provides no install instructions or homepage/source. Either the environment must already have a compatible `dbskiter` binary and configuration, or the skill cannot perform its stated purpose. This is an incoherence: a database inspector legitimately needs a client/binary and connection details, which are not declared.

Instruction Scope

Runtime instructions tell the agent to execute commands against databases (run, report, baseline, intelligent, anomalies, root-cause, risks). They do not explain how the agent obtains database connection info, credentials, or metrics (no env vars, config paths, or prompts described). The instructions therefore implicitly require access to sensitive secrets and monitoring data but provide no guidance on where those are sourced or how to limit/explain data flows. There is no direction to avoid reading unrelated files, but the lack of explicit authentication steps is a gap.

Install Mechanism

This is an instruction-only skill with no install spec and no code files. That minimizes the risk of arbitrary code being written to disk by the skill itself. However, it increases dependence on the runtime environment having the expected tooling (dbskiter) preinstalled.

Credentials

The skill declares no required environment variables or credentials, yet performing database inspections normally requires DB credentials, monitoring API keys, or access to metric stores. The absence of declared secrets or config paths is disproportionate to the described functionality and leaves open questions about how sensitive connection data will be supplied or used.

Persistence & Privilege

The skill is not forced-always and does not request special persistence. It uses the platform-default model invocation behavior. There is no install-time persistence or modification of other skills/configurations indicated.

What to consider before installing

This skill appears to be only usage instructions for a CLI named `dbskiter`, but it doesn't say where that CLI comes from or how it will authenticate to target databases. Before installing or invoking it: 1) Confirm you have a trusted `dbskiter` binary (source, version, and install steps). 2) Ask the author how DB credentials or metric data are supplied (explicit env vars, config files, or interactive prompts). 3) Do not let the agent run against production DBs until you verify credential scope (use a read-only/test account). 4) Prefer running this in an isolated environment and request the skill metadata be updated to declare required binaries and credential/env requirements. If the author cannot clarify these points, treat the skill as untrusted.

数据库巡检 Skill

何时使用

当用户提到以下关键词时，使用此skill：

用户说法	执行命令	说明
"巡检"	dbskiter --output-mode=ai --database=<name> inspector run	执行完整巡检
"生成报告"	dbskiter --output-mode=ai --database=<name> inspector report	生成巡检报告
"检查配置"	dbskiter --output-mode=ai --database=<name> inspector run --type configuration	仅检查配置
"检查安全"	dbskiter --output-mode=ai --database=<name> inspector run --type security	仅安全检查
"建立基线"	dbskiter --output-mode=ai --database=<name> inspector baseline --create	创建性能基线
"智能巡检"	dbskiter --output-mode=ai --database=<name> inspector intelligent	智能巡检分析
"异常检测"	dbskiter --output-mode=ai --database=<name> inspector anomalies	检测指标异常
"根因分析"	dbskiter --output-mode=ai --database=<name> inspector root-cause	分析问题根因
"风险预测"	dbskiter --output-mode=ai --database=<name> inspector risks	预测未来风险

核心命令（7个）

1. 执行完整巡检

dbskiter --database=<数据库名> inspector run

输出：健康评分、风险统计、巡检项详情

评分标准：

• 90-100：优秀
• 70-89：良好
• <70：需要关注

2. 生成报告

dbskiter --database=<数据库名> inspector report --output report.html

支持格式：HTML、Markdown、JSON

3. 基线管理

dbskiter --database=<数据库名> inspector baseline --create

用途：建立性能基线，用于后续对比

4. 智能巡检（新增）

dbskiter --database=<数据库名> inspector intelligent --metrics-history='{"cpu": [...], "memory": [...]}'

功能：异常检测、根因分析、风险预测、智能建议

5. 异常检测（新增）

dbskiter --database=<数据库名> inspector detect-anomalies --metrics='{"cpu": [...]}'

功能：检测突然飙升、逐渐增长、周期性异常、基线偏离

6. 根因分析（新增）

dbskiter --database=<数据库名> inspector analyze-root-causes --anomalies='[...]'

功能：分析异常事件的根因，提供解决建议

7. 风险预测（新增）

dbskiter --database=<数据库名> inspector predict-risks --time-horizon=7d

功能：预测未来7天/30天的容量和性能风险

8. 智能建议（新增）

dbskiter --database=<数据库名> inspector smart-recommendations

功能：基于巡检结果生成优先级排序的优化建议

9. 关联分析（新增）

dbskiter --database=<数据库名> inspector analyze-correlations --metrics='{"cpu": [...], "memory": [...]}'

功能：分析指标间关联性，发现隐藏模式

巡检类型

• configuration：配置检查
• performance：性能检查
• storage：存储检查
• security：安全检查
• capacity：容量检查

AI决策流程

场景1：用户说"巡检一下数据库"

步骤1：执行 dbskiter --database=<name> inspector run
步骤2：解读健康评分和风险统计
步骤3：如果有严重问题，建议进一步诊断
步骤4：生成报告：dbskiter --database=<name> inspector report

场景2：用户说"分析一下为什么CPU高"

步骤1：执行 dbskiter --database=<name> inspector root-cause --issue="CPU飙升"
步骤2：查看根因分析结果
步骤3：给出解决建议

场景3：用户说"预测一下未来风险"

步骤1：执行 dbskiter --database=<name> inspector risks --days=30
步骤2：解读风险预测结果
步骤3：给出预防和应对措施