dino-manage-boxes

Pass

Audited by Gen Agent Trust Hub on Oct 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data such as box paths, descriptions, and colors which are subsequently passed as arguments to the dino CLI.
  • Ingestion points: User-provided values for paths and descriptions in commands like dino box add, rename, and merge as specified in SKILL.md.
  • Boundary markers: The 'Runtime Contract' in SKILL.md explicitly instructs the agent to keep every user-provided value as a separate argument and never interpolate untrusted text into shell source.
  • Capability inventory: The skill utilizes the dino CLI binary for managing card box hierarchies.
  • Sanitization: Instructions mandate treating all user input and CLI output as untrusted, and require a dry-run/preview workflow for all destructive or additive mutations.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 31, 2026, 11:53 AM