The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes Node.js scripts (getnote.mjs, oauth-poll.mjs, common.mjs) to perform API tasks. It defaults to a preview mode and requires the --execute flag for live operations, ensuring user control over all actions.
[EXTERNAL_DOWNLOADS]: Interacts with the official Get 笔记 API at openapi.biji.com and performs image uploads to service-specified storage. These network operations are strictly limited to the intended service provider.
[CREDENTIALS_UNSAFE]: Handles API keys using a local auth.json file stored in the user's home directory (~/.config/flc1125/skills/getnote/). The skill correctly redacts these keys from console output and provides clear guidance on secure configuration.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from note content and link summaries, which constitutes a potential ingestion surface.
Ingestion points: Note content fields and link metadata retrieved during the link-saving and searching workflows in scripts/getnote.mjs.
Boundary markers: No specific delimiters or safety instructions are embedded in the data processing prompts.
Capability inventory: The skill possesses network communication capabilities, local file read access for images, and local file write access for authentication setup.
Sanitization: Employs a sanitizeJsonText function in scripts/common.mjs to prevent malformed data from affecting JSON parsing.

getnote