moai-domain-database
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (MEDIUM): The skill metadata in
SKILL.mdallows the use ofBash(npm:*)andBash(npx:*). This configuration permits the agent to download external packages from the npm registry. - REMOTE_CODE_EXECUTION (MEDIUM): Permission to use
npxallows for the direct execution of remote scripts and binaries, representing a significant risk if the agent is manipulated into running untrusted third-party tools. - COMMAND_EXECUTION (LOW): The skill allows access to several database CLI tools (
psql,mysql,sqlite3,mongosh,redis-cli,prisma). While restricted to database operations, these tools provide a broad interface for interacting with database servers and local environments. - PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8).
- Ingestion points: Untrusted data is ingested from PostgreSQL, MongoDB, Oracle, and Redis query results as described in the
modules/directory. - Boundary markers: Absent; there are no instructions or delimiters to distinguish between data and commands within the database outputs.
- Capability inventory: The agent has high permissions including
Bash(for database and package tools),Write, andEditfile system access. - Sanitization: Absent; the skill does not specify any methods for escaping or validating database content before processing.
Audit Metadata