realtime-monitor
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of Markdown instructions and JSON tool call templates. It does not contain any Python, JavaScript, or shell scripts, eliminating the risk of direct code execution.
- [DATA_EXPOSURE] (SAFE): Analysis of the skill body shows no hardcoded credentials, API keys, or attempts to access sensitive local file paths like SSH keys or environment variables.
- [REMOTE_CODE_EXECUTION] (SAFE): There are no patterns involving the download of external scripts (e.g., curl | bash) or the use of package managers to install untrusted dependencies.
- [INDIRECT_PROMPT_INJECTION] (SAFE): The skill is designed to process external stock market data. Although this is a theoretical ingestion point for untrusted content, the skill has no dangerous capabilities (such as file writing or command execution) that could be exploited if malicious data were returned by the API.
- Ingestion points: Tool outputs from the
stock-sdkMCP server. - Boundary markers: None explicitly defined in the prompt interpolation.
- Capability inventory: None; the skill is limited to displaying information.
- Sanitization: None; relies on the LLM's internal safety filters for data presentation.
- [PERSISTENCE_MECHANISMS] (SAFE): The provided YAML configuration for 'OpenClaw' is an illustrative example for a separate scheduling tool and does not represent an attempt to modify system crontabs or startup services on the agent's host.
Audit Metadata