PostgreSQL Database

PostgreSQL 数据库操作技能。用于执行SQL查询、表管理、备份恢复、性能监控等数据库操作。触发条件：用户提到 PostgreSQL、postgres、数据库查询、SQL查询、表结构、数据库备份等。

This skill appears to implement the PostgreSQL features it claims, but before installing: (1) note the registry metadata omits required PG environment variables — the scripts expect PGHOST/PGPORT/PGDATABASE/PGUSER/PGPASSWORD, so supply only a least-privileged DB user. (2) Ensure pg_dump/pg_restore and psycopg2 are installed on the host; the scripts use subprocess to call pg_dump and will fail if absent. (3) Backups are written to disk (backup_dir) — verify the directory, permissions, and retention policy so backups are not exposed. (4) Review the scripts yourself if you don't fully trust the source: they execute arbitrary SQL and run pg_restore/pg_dump, which can modify or drop data when run with sufficient privileges. (5) If you plan to allow the agent to invoke autonomously, restrict credentials and consider creating a dedicated read/backup role (avoid superuser) to limit blast radius.

Type: OpenClaw Skill Name: postgres-db Version: 1.0.0 The skill is classified as suspicious due to several critical vulnerabilities. The `scripts/query.py` file directly executes arbitrary SQL queries provided as command-line arguments, creating a significant SQL injection risk if an AI agent constructs these queries from untrusted user input. Additionally, `scripts/backup.py` executes external `pg_dump` and `pg_restore` commands with arguments derived from user input, which could lead to argument injection or unintended file operations. The `scripts/schema_export.py` script also presents a file write vulnerability by allowing the output path to be specified arbitrarily, potentially overwriting sensitive files. While these are not explicitly malicious actions, they are severe flaws that could be exploited for unauthorized database access, data manipulation, or system compromise.