connect-query

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill incorporates a shell script (scripts/connect-query-check.sh) designed to operate as a local git or IDE hook. The script uses standard tools like grep and cat to analyze the contents of TypeScript files during the development workflow.
  • [INDIRECT_PROMPT_INJECTION]: The linting script processes user-generated code (untrusted input) to identify and block patterns that violate project standards. It includes a documented escape hatch (// allow: direct-query) allowing developers to bypass checks for specific files when justified, which is a standard feature for linting tools.
  • [DYNAMIC_EXECUTION]: The shell script utilizes source to load a helper library (_hook-lib.sh). This is a typical modular design for local shell scripts and is used to provide common hook functionality.
  • [SAFE]: The skill's instructions and references align with established best practices for the ConnectRPC and Protobuf ecosystems. It references legitimate, well-known libraries such as @connectrpc/connect-query and @bufbuild/protobuf for their intended purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 02:00 AM
Security Audit — agent-trust-hub — connect-query