The Agent Skills Directory

External Content Retrieval: The skill fetches lesson content from the vercel.com domain to drive the guided learning loop. While this involves processing external markdown files, the source is an official platform associated with the skill's purpose.
Indirect Prompt Injection Surface: There is a potential surface for indirect prompt injection because the skill ingests untrusted markdown data from external URLs and includes capabilities to execute project setup commands.
Ingestion points: Lesson content is fetched from https://vercel.com/academy/slack-agents/*.md (defined in SKILL.md).
Boundary markers: No explicit boundary markers or 'ignore embedded instructions' warnings are present in the provided instructions for the fetched content.
Capability inventory: The skill utilizes shell commands including pnpm install, pnpm dlx, and various Vercel and Slack CLI commands (SKILL.md).
Sanitization: No explicit sanitization or validation of the fetched markdown content is described.
Project Scaffolding and Command Execution: The skill guides users through cloning a template repository and running installation commands like pnpm install. These are routine developer workflows for project initialization but involve executing code from a remote template.
Environment Secret Management: The instructions direct the user to store sensitive credentials (e.g., SLACK_BOT_TOKEN, AI_GATEWAY_API_KEY) in a .env file and use standard CLI tools for environment configuration. This follows industry best practices for local secret management.

slack-agents