dbvr
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
dbvrcommand-line interface to interact with databases, allowing the agent to execute SQL queries and browse schema metadata. This involves running subprocesses with user-controlled SQL strings and connection parameters. - [INDIRECT_PROMPT_INJECTION]: The skill processes data returned from database queries, which is a potential surface for indirect prompt injection if the database content contains malicious instructions.
- Ingestion points: Database query results from
dbvr sqland metadata listings fromdbvr meta(SKILL.md, references/command-reference.md). - Boundary markers: Absent; results are displayed as Markdown tables or plain lists without specific encapsulation or ignore-instructions.
- Capability inventory: Execution of the
dbvrbinary and the ability to write query results to arbitrary file paths via the-outflag. - Sanitization: None specified; results are presented to the user or saved to disk as returned by the driver.
Audit Metadata