The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes documentation for processing Vercel API responses using tools like jq, creating a potential ingestion point for data within the agent's context. As the data source is the user's own authorized Vercel account and the functionality is central to the skill's management purpose, this is documented as a standard operational pattern without security risk.
Ingestion points: API response output in SKILL.md (Step 2 and Step 6).
Boundary markers: No specific delimiters are implemented in the shell examples.
Capability inventory: The skill is configured with Bash(curl:*), Read, Write, and Edit capabilities.
Sanitization: Data is structured via jq, though no explicit prompt injection sanitization is present in the examples.
[SAFE]: The skill utilizes environment variables such as $VERCEL_TOKEN and process.env.SAML_CERT for authentication, which is a secure and standard practice.
[SAFE]: External references to jeremylongshore.com, intentsolutions.io, and tonsofskills.com are legitimate vendor-owned resources that align with the skill's author context.

vercel-enterprise-rbac