sql-injection-stacked
Warn
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill facilitates remote code execution on target database servers by providing ready-to-use exploitation flows. It includes instructions for using MSSQL's
xp_cmdshell, PostgreSQL'sCOPY TO PROGRAM, Oracle'sDBMS_SCHEDULER, and custom libc function calls to execute arbitrary code on the host system. - [COMMAND_EXECUTION]: Includes specific examples of command-line execution, most notably a Base64-encoded PowerShell payload (
JABjAGwAaQBl...) which decodes to a functional TCP reverse shell. It also provides bash-based reverse shell commands for PostgreSQL environments. - [DATA_EXFILTRATION]: Provides methodologies for unauthorized data access and exfiltration, such as reading sensitive files like
/etc/passwdusing theCOPYcommand or creating temporary database tables to stage and extract sensitive application data. - [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it instructs the agent to ingest and process untrusted data from target database outputs and injection parameters. There are no boundary markers or sanitization instructions to mitigate malicious instructions embedded in the processed data. (Ingestion points: SQL parameter inputs, database query results; Capability inventory: OS command execution via xp_cmdshell/COPY, file system writes via INTO OUTFILE, tool execution with sqlmap; Sanitization: Absent; Boundary markers: Absent).
Audit Metadata