cte-query-builder

Pass

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [NO_CODE] (SAFE): The skill consists solely of markdown metadata and documentation with no associated scripts or configuration files.
  • [SAFE] (SAFE): Analysis of the markdown content revealed no signs of obfuscation, hardcoded credentials, or malicious prompt injection patterns.
  • [Indirect Prompt Injection] (LOW): The skill is designed to process untrusted user data for data analytics tasks.
  • Ingestion points: User requests involving 'cte query builder' patterns or data analytics questions.
  • Boundary markers: No delimiters or 'ignore' instructions are provided to separate user inputs from agent instructions.
  • Capability inventory: The skill authorizes the use of powerful tools including Bash, Write, Edit, Read, and Grep.
  • Sanitization: No sanitization or validation logic is defined to mitigate the risk of instructions embedded within user-provided schemas or requirements.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 18, 2026, 03:25 PM