universal-db-mcp-connector
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install the "universal-db-mcp" package from the npm registry and clone the source code from a third-party GitHub repository ("Anarkh-Lee/universal-db-mcp"). These sources are outside of verified organizational scopes.
- [REMOTE_CODE_EXECUTION]: The documentation encourages the use of "npx" to download and execute the "universal-db-mcp" tool. This pattern facilitates the execution of remote code from unverified third-party sources.
- [COMMAND_EXECUTION]: The documentation explicitly instructs the execution of shell commands to install software globally ("npm install -g"), build from source, and configure the environment.
- [INDIRECT_PROMPT_INJECTION]: The skill exposes the agent to data from external databases through tools like "execute_query" and "get_sample_data". This creates a vulnerability surface where malicious data stored in a database could influence the agent's logic or behavior.
- Ingestion points: SQL query results and table samples retrieved from connected databases.
- Boundary markers: No specific boundary markers or instructions to ignore embedded commands in database data are present in the provided instructions.
- Capability inventory: Includes arbitrary SQL execution ("execute_query"), dynamic database connection management ("connect_database"), and schema inspection tools.
- Sanitization: No evidence of data sanitization or filtering of database content before it is processed by the AI context.
- [METADATA_POISONING]: The skill claims to be authored by "ara.so", which differs from the internal author context assigned to this skill. This discrepancy could lead to confusion regarding the source and trustworthiness of the instructions.
Audit Metadata