scaffold-exercises

Pass

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires executing mkdir, git, and pnpm ai-hero-cli to scaffold files and validate their structure.
  • [INDIRECT_PROMPT_INJECTION]: The workflow involves parsing a plan to generate directory paths. Ingestion point: Step 1 (Parse the plan). Boundary markers: Absent. Capability inventory: mkdir, git, pnpm. Sanitization: Absent. This ingestion of untrusted data represents an attack surface for manipulating command arguments.
  • [DYNAMIC_EXECUTION]: The agent is instructed to dynamically build shell commands based on the input plan to create directories and move folders.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 28, 2026, 11:06 AM