recipe-simplify-codebase

Warn

Audited by Gen Agent Trust Hub on Jul 11, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands by interpolating variables like {recipe_thoroughness}, {recipe_scope}, {plan_id}, and {expanded_instruction} into python3 .plan/execute-script.py calls in Steps 0, 1, and 5. This pattern is susceptible to command injection if the input values (particularly plan_id or values returned by intermediate scripts) contain shell metacharacters.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting architectural metadata and user-supplied parameters to generate a 'Solution Outline' in Step 5.
  • Ingestion points: Architectural unit lists from manage-architecture:architecture modules (Step 2) and user inputs (Step 0).
  • Boundary markers: There are no identified delimiters or instructions to ignore embedded content within the generated outline template.
  • Capability inventory: The skill has the ability to execute shell commands via internal scripts and write to the file system using the Write tool.
  • Sanitization: No explicit sanitization or validation of the architectural data or user input is performed before interpolation into the final document.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 11, 2026, 07:12 AM
Security Audit — agent-trust-hub — recipe-simplify-codebase