skills/udecode/plate/generate_command/Gen Agent Trust Hub

generate_command

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the '#$ARGUMENTS' placeholder, which allows arbitrary user input to be embedded into the core logic of a generated command without sanitization.
  • Ingestion points: The '#$ARGUMENTS' variable in SKILL.md is used to capture user requirements.
  • Boundary markers: Absent; user-provided text is directly interpolated into the command goal.
  • Capability inventory: The generated commands are explicitly designed to leverage powerful capabilities including Bash execution, File Operations (Read/Write/Edit), and WebFetch.
  • Sanitization: Absent; there are no instructions or patterns provided to validate or escape the user input.
  • [COMMAND_EXECUTION]: The skill provides a framework for generating tools that execute arbitrary bash commands (e.g., 'bin/rails test', 'git commit'). While intended for development tasks like testing and linting, this provides a mechanism for system-level interaction and potential misuse if the generated instructions are not carefully reviewed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 07:17 AM