hidden-setting-does-not-clear-the-stored-value

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions are purely technical and educational, focused on Kotlin development practices. No attempts to override system prompts or bypass safety filters were detected.
  • [DATA_EXFILTRATION]: No network operations or access to sensitive file paths were identified. The included shell commands are restricted to searching local source files using grep.
  • [REMOTE_CODE_EXECUTION]: There are no patterns involving downloading and executing remote scripts. The commands provided are standard development tools for local codebase inspection.
  • [COMMAND_EXECUTION]: The skill uses grep for verifying code patterns within the src directory. These are safe, read-only operations typical for development workflows.
  • [OBFUSCATION]: The content is clear and uses standard Markdown and Kotlin code. No hidden characters, Base64 encoding, or steganographic techniques were found.
  • [INDIRECT_PROMPT_INJECTION]: While the skill discusses handling untrusted stored data (preferences), the logic it teaches is a defensive measure to ensure UI rendering stability. No exploitable surface for indirect injection was found.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 07:02 AM
Security Audit — agent-trust-hub — hidden-setting-does-not-clear-the-stored-value