Db Explorer

PassAudited by VirusTotal on May 10, 2026.

Overview

Type: OpenClaw Skill Name: db-explorer-tool Version: 2.0.0 The db-explorer skill (SKILL.md) provides an AI agent with comprehensive instructions to interact with various databases using CLI tools like psql, mysql, and mongosh. Although it includes explicit safety rules such as read-only defaults, query limits, and transaction wrapping, the skill facilitates high-risk behaviors including shell command execution, database credential handling, and local file system writes (e.g., to /tmp/export.csv). These capabilities are consistent with the tool's stated purpose but present a broad attack surface for potential misuse, justifying a suspicious classification.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If the user provides privileged credentials, the agent could query, export, or, with confirmation, modify important database data.

Why it was flagged

The skill requires database credentials or connection details. This is expected for a database explorer, but those credentials may grant access to sensitive or production data.

Skill content
Ask the user for:
- Database type (postgres/mysql/sqlite/mongo/redis)
- Connection string OR host/port/database/user/password
Recommendation

Use read-only or least-privilege credentials whenever possible, avoid production admin accounts, and do not share passwords beyond the specific database task.

What this means

A confirmed write or destructive query could permanently change or remove database records.

Why it was flagged

The skill contemplates write, delete, and drop operations against databases. It does include an explicit confirmation requirement, making this purpose-aligned but still high-impact.

Skill content
Never run INSERT/UPDATE/DELETE/DROP without explicit user confirmation
Recommendation

Review every write query before approving it, prefer transactions and rollback previews, and keep backups or recovery plans for important databases.

What this means

The agent could read or export more data than intended, or run a Redis command that affects performance on a large or production instance.

Why it was flagged

Some example commands are broad: exporting SELECT * can copy entire tables, and Redis KEYS "*" can be expensive on large Redis instances.

Skill content
psql "$CONN" -c "\copy (SELECT * FROM table_name) TO '/tmp/export.csv' WITH CSV HEADER" ... redis-cli -h host -p 6379 -a password KEYS "*"
Recommendation

Add WHERE clauses and LIMITs for exports, choose safe output paths, and use Redis SCAN instead of KEYS "*" on production-sized databases.

What this means

Installing external tools changes the local environment and depends on the trustworthiness of the package source.

Why it was flagged

The skill relies on external database command-line clients installed by the user. This is normal for the purpose, and there is no automatic install script in the provided artifacts.

Skill content
CLI Tool     | Install (macOS)           | Install (Linux) ... psql ... brew install postgresql ... apt install postgresql-client
Recommendation

Install database clients only from trusted package managers or official vendor documentation, and verify what will be installed.