The Agent Skills Directory

[SAFE]: The skill implements 44 tools for interacting with the Trello REST API, covering boards, lists, cards, labels, and checklists.
[SAFE]: Authentication is handled either via Zapier-managed OAuth or direct environment variables (TRELLO_API_KEY and TRELLO_TOKEN), following secure secret management practices.
[SAFE]: Tools that fetch remote files for attachments correctly use globalThis.fetch in scripts/addCardAttachment.ts and scripts/createCard.ts to prevent leaking Trello API credentials to third-party servers when processing user-provided URLs.
[SAFE]: The cli.js entry point uses spawnSync to execute its own source, facilitating platform compatibility without executing remote code.
[SAFE]: The skill includes an indirect prompt injection surface by ingesting Trello card data.
Ingestion points: Card names, descriptions, and comments are read in listCards.ts, searchCards.ts, and getCard.ts.
Boundary markers: Zod schemas validate data structure, though content is processed as raw strings.
Capability inventory: The skill can perform network operations via ctx.fetch and modify Trello resources.
Sanitization: All external content used in API requests is validated via Zod and properly URL-encoded.

trello