The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the node:child_process module to invoke external tools like java (for the TLA+ model checker), mise (for tool management), and JavaScript runtimes.
To enhance security, the childEnv function in scripts/run.ts restricts the environment variables passed to child processes, ensuring that sensitive information such as proxy credentials or internal tokens are not leaked to external tools.
[EXTERNAL_DOWNLOADS]: The skill fetches the TLA+ toolset (tla2tools.jar) from the official Tlaplus organization's GitHub repository. This download is conducted via the mise tool and targets a well-known, trusted source.
[DYNAMIC_EXECUTION]: The skill supports an extensible visualization architecture where users can provide their own JavaScript modules via a --view flag. These modules are imported at runtime to customize how state labels and transitions are described in the resulting statechart.
[OBFUSCATION]: The system implements a comprehensive escaping and sanitization layer in scripts/backends/mermaid.ts. This logic uses specific character sequences (e.g., ﬂ°°, ¶ß) as internal markers to ensure that model data and user-provided strings cannot break out of the Mermaid syntax or execute unauthorized Mermaid directives.

statechart