Aloudata CAN SKILLS - text-to-sql-query
Security checks across static analysis, malware telemetry, and agentic risk
Overview
The skill is coherent for direct retail database querying, but it uses an API key to run generated SQL and can return sensitive business or customer-related data, so permissions should be reviewed.
Install this only in an environment where CAN_API_KEY is a least-privilege, read-only key for the intended retail tables. Users should review or constrain generated SQL, especially for member/customer-related data, and avoid broad result sets unless they are authorized and necessary.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
A mistaken or overly broad natural-language request could cause the agent to retrieve more retail data than the user intended.
The skill explicitly has the agent generate and execute SQL against a database. This is central to the purpose and mitigated by documented SELECT/table/LIMIT restrictions, but generated SQL can still retrieve unintended data if the prompt or query is too broad.
根据用户自然语言描述，直接生成 SQL 查询语句，通过 Gateway JDBC SQL 直查接口在零售数据库上执行并返回结果。
Review generated SQL for sensitive columns, filters, and row limits before relying on results; use least-privilege database access where possible.
Anyone or any agent session with access to CAN_API_KEY may be able to query the allowed retail database tables through this skill.
The skill requires an API key to access the database gateway. That is expected for this integration and is not hardcoded, but the provided registry requirements list no required environment variables or primary credential.
API Key 通过环境变量 `$CAN_API_KEY` 注入，禁止在 Skill 文件中硬编码。
Declare the credential requirement in metadata, scope the API key to read-only access for the intended tables, and rotate or revoke it when no longer needed.
Sensitive query results could appear in the agent conversation or be exposed to users who should not see those rows or fields.
The skill sends generated SQL to an external gateway and returns query results to the agent. This data flow is disclosed and purpose-aligned, but query results may include sensitive business or member-related records.
POST https://gateway.can.aloudata.com/api/jdbc/query ... 请求体：{"sql": "SELECT ... FROM table_name WHERE ... LIMIT N"}Avoid selecting unnecessary identifiers or member-level data, apply filters and limits, and ensure the gateway enforces the organization's access-control and logging policies.