graphql-architect

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • PROMPT_INJECTION (SAFE): The instructions are focused on GraphQL design and optimization. There are no attempts to bypass safety filters, extract system prompts, or override agent constraints.
  • DATA_EXFILTRATION (SAFE): No commands or instructions for network access or reading sensitive local files were found.
  • EXTERNAL_DOWNLOADS (SAFE): No external dependencies or remote script downloads are specified.
  • COMMAND_EXECUTION (SAFE): There are no shell commands, script executions, or subprocess calls defined in the skill.
  • CREDENTIALS_UNSAFE (SAFE): No hardcoded secrets, API keys, or private tokens are present in the metadata or body.
  • INDIRECT_PROMPT_INJECTION (SAFE): While the skill is designed to process user-provided GraphQL requirements, it lacks any automated tool execution capabilities that would create a high-risk attack surface.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:33 PM