mental-model-mapper
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No evidence of malicious instructions, safety bypasses, or jailbreak attempts. The instructions focus entirely on mapping mental models using established systems-thinking frameworks.
- [DATA_EXFILTRATION]: No network operations, API calls, or file system access patterns were detected. The skill operates solely on the provided user inputs within the prompt context.
- [REMOTE_CODE_EXECUTION]: The skill does not include or reference any external scripts, binaries, or package installations.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided text through fields like
system_focusandvisible_evidence. While it lacks explicit boundary markers (e.g., XML tags or delimiters) to isolate this input, the skill does not possess any dangerous capabilities (such as shell access, network requests, or file writes) that could be exploited via injection. The risk is limited to potential manipulation of the generated text output. - [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or credentials were found in the skill metadata or instructions.
- [COMMAND_EXECUTION]: No shell commands, subprocess spawning, or dynamic context injection patterns (
!command) are present.
Audit Metadata