The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documents the Neutralino.updater.checkForUpdates(url) API and the neu create --template <account>/<repo> command, both of which involve fetching resources or project code from remote locations.
[COMMAND_EXECUTION]: Provides instructions and examples for system-level operations using Neutralino.os.execCommand and Neutralino.os.spawnProcess, which are core features of the framework intended for process management.
[DYNAMIC_EXECUTION]: Describes the application's ability to update its own resource bundles at runtime using the Neutralino.updater.install() method.
[INDIRECT_PROMPT_INJECTION]: The skill documents an framework architecture with a significant attack surface for indirect prompt injection.
Ingestion points: External data enters the context through filesystem.readFile, clipboard.readText, and the updater API.
Boundary markers: Security boundaries are defined using nativeAllowList, nativeBlockList, and tokenSecurity (one-time token) settings.
Capability inventory: The framework provides extensive system capabilities including os.execCommand, os.spawnProcess, filesystem.writeFile, and filesystem.setPermissions.
Sanitization: The provided documentation does not specify sanitization or escaping requirements for data ingested from untrusted external sources.

neutralinojs