yes-and

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [NO_CODE]: The skill is entirely composed of Markdown instructions and a prompt template. It does not include any scripts, binaries, or configuration files that execute code.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it is designed to process untrusted user-supplied data within a prompt template.
  • Ingestion points: The {{PASTE THE IDEA...}} placeholder in SKILL.md allows arbitrary external text into the context.
  • Boundary markers: Present. The skill uses <context> tags to isolate the user-provided input from the system instructions.
  • Capability inventory: None. The skill manifest does not define any allowed-tools or scripts, meaning it cannot perform network requests, file operations, or command execution.
  • Sanitization: Not present. The skill relies on the model's internal safety filters and the provided XML-style delimiters.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 08:18 PM
Security Audit — agent-trust-hub — yes-and