thinking-ray-dalio

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The content consists entirely of instructional advice and decision-making frameworks. There are no attempts to override agent behavior, bypass safety filters, or use 'Ignore previous instructions' patterns.
  • [EXTERNAL_DOWNLOADS]: No external network requests, URL fetches, or remote downloads are present in the skill.
  • [COMMAND_EXECUTION]: The skill does not contain any shell commands, subprocess calls, or scripts. It is purely markdown-based documentation.
  • [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or secrets were detected. The mentions of internal tools like 'Dot Collector' are descriptive examples with no associated access credentials.
  • [DATA_EXFILTRATION]: There are no mechanisms for accessing sensitive system files or transmitting data to external domains.
  • [REMOTE_CODE_EXECUTION]: No patterns for downloading or executing remote code or scripts (e.g., curl|bash) were found.
  • [DYNAMIC_EXECUTION]: The skill does not use runtime code generation, unsafe deserialization, or dynamic path loading.
  • [OBFUSCATION]: The text was analyzed for Base64, zero-width characters, homoglyphs, and hidden URLs; no such techniques were found.
  • [SAFE]: The skill follows best practices for instructional content, providing templates for user input without exploitable capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 09:29 AM
Security Audit — agent-trust-hub — thinking-ray-dalio