The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install better-auth and its related ecosystem packages (e.g., @better-auth/cli, @better-auth/passkey) via npm. These are legitimate, well-known packages for the intended functionality.
[COMMAND_EXECUTION]: The skill utilizes command execution for environment setup and database management. Evidence includes:
Installation commands: npm install better-auth.
Migration and schema generation: npx @better-auth/cli@latest migrate and npx @better-auth/cli@latest generate.
Secret generation: Recommends using openssl rand -base64 32 for creating secrets.
[INDIRECT_PROMPT_INJECTION]: The skill scans existing project files to detect frameworks and databases, creating a potential surface for indirect injection if a malicious file contains instructions designed to influence the agent's behavior during Phase 1.
Ingestion points: Project configuration files (next.config, package.json, prisma/schema.prisma) are read in Phase 1, Step 1.
Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the file scanning process.
Capability inventory: The agent has the capability to execute shell commands for package installation and CLI operations (npm, npx).
Sanitization: No specific sanitization or validation of the content of the scanned files is specified prior to logic branching.

create-auth-skill