querypie

Fail

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides an instruction to download and execute a shell script from a remote URL (curl -fsSL ... | sh) if the querypie tool is missing. This executes unverified code from a non-pre-vetted source directly on the host system.
  • [COMMAND_EXECUTION]: The skill is designed to execute various shell commands through the querypie CLI to list databases, schemas, and perform SQL queries.
  • [EXTERNAL_DOWNLOADS]: The skill fetches an installation script from a remote GitHub repository (raw.githubusercontent.com/sudosubin/querypie-cli/main/scripts/install.sh) during the setup phase.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface when processing data returned from database queries.
  • Ingestion points: Results from SQL queries and schema descriptions retrieved via the CLI.
  • Boundary markers: Absent; there are no delimiters or instructions to ignore potential commands embedded in database content.
  • Capability inventory: The agent can execute shell commands and write to the local filesystem using the provided toolset.
  • Sanitization: Absent; query results are processed as plain text without validation or filtering.
Recommendations
  • HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/sudosubin/querypie-cli/main/scripts/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 17, 2026, 03:48 PM
Security Audit — agent-trust-hub — querypie