chalk-logic

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues were identified during the analysis of the skill. The logic is focused on transforming user concepts into detailed visual prompts for image generation.
  • [PROMPT_INJECTION]: The instructions do not contain patterns designed to bypass safety filters or override agent behavior. The process is highly constrained by a "Visual contract" and specific recipe axes.
  • [DATA_EXFILTRATION]: There are no network operations or access to sensitive local files. The skill uses local reference images for visual anchoring only.
  • [REMOTE_CODE_EXECUTION]: The skill consists entirely of markdown instructions and YAML configuration. No code execution, script generation, or external dependencies are present.
  • [COMMAND_EXECUTION]: The skill does not use shell commands or perform any system-level operations. Instructions regarding "isolated workers" and "subagents" describe internal agent workflow management rather than external execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 02:38 PM
Security Audit — agent-trust-hub — chalk-logic