The Agent Skills Directory

[COMMAND_EXECUTION]: The cli.js entry point uses node:child_process.spawnSync to execute the connector's internal logic. This is a standard part of the skill's architecture for managing its own execution environment.
[EXTERNAL_DOWNLOADS]: The addCardAttachment and createCard tools support downloading files from remote URLs to be uploaded as Trello attachments. The implementation correctly uses a separate fetch mechanism to ensure Trello authentication tokens are never sent to external servers.
[DATA_EXFILTRATION]: The skill is designed to interact with the Trello API and user-provided URLs. The code demonstrates defensive programming by ensuring sensitive Trello credentials are only shared with the official Trello API endpoints and not leaked during external resource requests.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes data from Trello cards and comments.
Ingestion points: Card names, descriptions, and comments retrieved via tools like getCard, listCards, and searchCards.
Boundary markers: None explicitly defined within the scripts.
Capability inventory: Tool execution including Trello API writes and external file downloads.
Sanitization: Input validation via Zod schemas and character limit slicing for descriptions.

trello