Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计异常

postgres-cliPostgres CLI 搜索

Agent Skill

用于辅助数据库表结构、查询语句、迁移脚本和数据维护任务。它适合让 Agent 分析 schema、编写 SQL、排查查询问题、整理索引或生成迁移建议。使用时需要明确数据库类型、连接环境和目标表,区分只读分析与写入变更;涉及删除、更新、迁移和批量导入时,应优先 dry-run、备份或事务保护,避免误操作。

总安装

396

周安装

16

GitHub Stars

1

下载量

124
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:postgres-cli(Postgres CLI 搜索)
来源仓库:https://github.com/dobby/postgres-cli-skill
仓库路径:skills/postgres-cli
安装命令:
npx skills add https://github.com/dobby/postgres-cli-skill --skill postgres-cli
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/dobby/postgres-cli-skill --skill postgres-cli

简介

用于辅助数据库表结构、查询语句、迁移脚本和数据维护任务。

  • 适合分析 schema、编写 SQL、排查查询问题或生成索引优化建议。
  • 使用时需明确数据库类型和连接环境,区分只读分析与写入操作。
  • 涉及删除、更新、迁移或批量导入时,应优先 dry-run、备份或使用事务保护。
  • 建议在安装前确认数据库兼容性,并验证技能是否依赖特定 CLI 工具或网络资源。

SKILL.md

Postgres CLI (V2)

Use postgres-cli to query PostgreSQL through named connection targets.

Platform Support

  • scripts/postgres-cli is the canonical launcher path inside the installed skill directory.
  • Prebuilt binaries are expected in scripts/bin/ for:

- macOS arm64 (postgres-cli-darwin-arm64) - Linux x86_64 (postgres-cli-linux-x86_64) - Windows x86_64 (postgres-cli-windows-x86_64.exe)

  • If no compatible binary exists and source + Cargo are available, launcher falls back to cargo run --release.

Available Scripts

  • scripts/postgres-cli - Launcher that selects a platform binary.
  • scripts/build-release-binary.sh - Builds and places a binary for current host into scripts/bin/.
  • scripts/refresh-binaries-from-release.sh <tag> - Downloads release binaries into scripts/bin/.

When To Use

  • The user asks to inspect Postgres data.
  • The user asks to run SQL against a configured target.
  • The user asks for schema introspection.
  • The user asks to validate DB CLI config or debug connectivity.
  • The user asks to refresh schema cache for agent context.

Setup

Read Setup Guide.

V2 Command Contract

Global flags (before subcommand):

  • --project-root <path> (optional; default cwd)
  • --target <name> (optional if default_target set)
  • --format <json|text|csv|tsv> (default json)
  • --output <path> (optional output file)
  • --no-summary (suppresses text summary)

Subcommands:

  • query

- input: exactly one of --sql, --sql-file, --stdin - flags: --mode <read|write> (default read), --timeout-ms

  • explain

- input: exactly one of --sql, --sql-file, --stdin - flags: --analyze, --verbose, --buffers, --settings, --wal, --timeout-ms

  • introspect

- required: --kind <schemas|tables|columns|indexes|constraints|views|materialized-views|functions|triggers|enums|rowcounts|rowcounts-exact> - optional filters: --schema (repeatable), --table schema.table (repeatable)

  • schema-cache update

- flags: --all-tables, --with-markdown, --table-file-naming <table|schema-table>, --timeout-ms

  • targets list
  • config validate
  • doctor

Safety Rules

  • Prefer read targets for normal data inspection.
  • query --mode read blocks mutating SQL.
  • Mutating SQL requires query --mode write and allow_write=true on target.
  • explain --analyze on mutating SQL requires write-enabled target.

Operational Guardrails

  • MUST use scripts/postgres-cli from the installed skill directory for all database interactions.
  • MUST NOT execute psql directly, even if psql is installed.
  • MUST NOT read secret/config files directly:

- .agents/postgres-cli/postgres.toml - .agent/postgres-cli/postgres.toml - .agents/postgres-cli/.env - .agent/postgres-cli/.env - .env at repository root

  • MUST treat target selection as:

- If the user provides a connection name, pass --target <name>. - If the user does not provide a connection, omit --target and rely on CLI fallback to default_target. - If the CLI returns TARGET_MISSING, ask the user for a connection name and do not inspect config files.

  • MUST NOT bypass CLI write-safety controls by using direct DB tools.

Command Patterns

Run read query:

scripts/postgres-cli --project-root /path/to/repo --target app-read query --sql "SELECT now();"

Run write query intentionally:

scripts/postgres-cli --project-root /path/to/repo --target app-write query --mode write --sql "UPDATE users SET active = true WHERE id = 1;"

Introspect tables:

scripts/postgres-cli --project-root /path/to/repo --target app-read introspect --kind tables

Explain a query:

scripts/postgres-cli --project-root /path/to/repo --target app-read explain --sql "SELECT * FROM users WHERE id = 1;"

Validate config:

scripts/postgres-cli --project-root /path/to/repo config validate

Doctor connectivity:

scripts/postgres-cli --project-root /path/to/repo --target app-read doctor

Update schema cache (JSON only):

scripts/postgres-cli --project-root /path/to/repo --target app-read schema-cache update

Update schema cache with markdown:

scripts/postgres-cli --project-root /path/to/repo --target app-read schema-cache update --with-markdown

Progressive Schema Loading

When schema context is needed, use this order:

  1. Read .agents/postgres-cli/schema/index.json.
  2. Load only required files from .agents/postgres-cli/schema/tables/*.json.
  3. Read .agents/postgres-cli/schema/relations.json when join/relationship reasoning is needed.
  4. If markdown was generated, consult .md files only for human-friendly display.
  5. If cache is missing or stale, run schema-cache update.

Agent Guidelines

  • Pass --target whenever the user provides a connection name.
  • If the user did not provide a connection name, allow default_target fallback.
  • Default to --format json for machine parsing.
  • Prefer query --mode read unless the user explicitly requests data mutation.
  • Return relevant result rows and summarize key findings.
  • If a consuming repo provides a repo-root scripts/postgres-cli wrapper, prefer that path to avoid launcher discovery work.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.63%
按下载量换算43

Claude

30.17%
按下载量换算37

Cursor

18.62%
按下载量换算23

Gemini CLI

8.83%
按下载量换算11

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills