The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates by executing the getnote CLI. Instructions include specific safeguards such as using --stdin or --content-file for processing potentially large or complex user-provided text to prevent shell injection or data truncation errors. Commands are restricted to the official binary defined in the skill metadata.
[EXTERNAL_DOWNLOADS]: The skill specifies a dependency on the @getnote/cli package via the platform-native install configuration in SKILL.md. This ensures that the installation of external tools is governed by the platform's security policies and requires explicit user consent during the skill installation process.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests content from external URLs and existing notes. The skill mitigates this risk through instructional design, such as requiring explicit user requests before expanding full note content in shared sessions and using structured input methods for data processing.
[CREDENTIALS_SAFE]: Authentication is handled via a secure login flow that directs the user to a browser-based authorization page managed by the official service provider. The skill instructions explicitly forbid requesting or storing sensitive tokens, API keys, or cookies within the agent's dialogue context.

getnote