The Agent Skills Directory

Management of Infrastructure via CLI: The skill utilizes the vercel firewall command suite to modify project security settings.
Context: These commands are used to create, edit, and delete firewall rules, IP blocks, and system bypasses. This is the primary purpose of the skill and is handled via the official platform CLI.
Mitigation: The skill reinforces a "best practice" workflow where changes are staged as drafts and requires the user to manually execute vercel firewall publish to apply changes, providing a human-in-the-loop safety check.
Analysis of Local Repository Content: The instructions suggest that the agent proactively scan the repository for API endpoints to recommend rate-limiting rules.
Security Consideration: This represents a surface for indirect prompt injection, as the agent ingests data from the codebase to influence its recommendations (Category 8).
Evidence Chain:
Ingestion points: Local repository files are scanned for API endpoint definitions.
Boundary markers: None specified for the scanning process.
Capability inventory: The skill can stage modifications to the project's firewall via vercel firewall rules add in SKILL.md.
Sanitization: Not explicitly mentioned, though recommendations are presented to the user for review.
Context: This is a standard developer assistant feature and is used here to improve the security posture of the application by identifying unprotected endpoints.

vercel-firewall