exploring-bitwarden-data

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on executing database client CLI tools to perform its primary function.
  • SKILL.md and the provider reference files (references/providers/mssql.md, references/providers/mysql.md, references/providers/postgresql.md) instruct the agent to use sqlcmd, mysql, psql, and docker exec to run queries.
  • The skill specifies mandatory flags (e.g., -K ReadOnly for sqlcmd) to restrict these commands to read-only operations.
  • [DATA_EXFILTRATION]: The skill processes sensitive database content and credentials, but implements strong defensive measures.
  • SKILL.md includes a "Secrets" section explicitly forbidding the logging or exporting of password environment variables.
  • It enforces driver-level or session-level read-only constraints (e.g., SET SESSION TRANSACTION READ ONLY;) as defense-in-depth.
  • [CREDENTIALS_UNSAFE]: Database authentication is managed via sensitive environment variables such as BW_MSSQL_PASSWORD and BW_POSTGRES_PASSWORD.
  • The skill follows security best practices by instructing that these credentials be passed inline to command-line tools rather than being exported globally or stored in cleartext files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 08:42 PM
Security Audit — agent-trust-hub — exploring-bitwarden-data