book-author-bio

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is designed purely for content generation and professional writing tasks. It provides templates and logical frameworks for producing author-related documentation without requesting dangerous system permissions or executing external code.
  • [EXTERNAL_DOWNLOADS]: The skill references several external URLs (KPIPA, Kyobo Book, Aladin, Korean Writers Association). These are all well-known, official Korean institutions and commercial bookstores used for research and style reference. No unauthorized or suspicious remote code downloads are present.
  • [PROMPT_INJECTION]: The instructions contain standard instructional keywords to guide the AI's behavior. There are no attempts to bypass safety filters, extract system prompts, or override the core agent's constraints.
  • [DATA_EXFILTRATION]: No evidence of hardcoded credentials, sensitive file path access, or network exfiltration patterns. The workflow involves processing user-provided information to generate text output within the chat context.
  • [COMMAND_EXECUTION]: The skill does not contain any shell commands, subprocess calls, or dynamic execution patterns. It relies entirely on the language model's text generation capabilities.
  • [REMOTE_CODE_EXECUTION]: There are no package installations (pip, npm) or remote script execution commands detected in the skill files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 06:27 AM
Security Audit — agent-trust-hub — book-author-bio