The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses a bash script (scripts/check-upstream.sh) to download markdown files and documentation from external GitHub repositories (robelest/convex-auth and get-convex/self-hosting) to the /tmp directory.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface detected.
Ingestion points: The agent is instructed to fetch and read external content from GitHub (e.g., README.md, INTEGRATION.md) and the website auth.estifanos.com to guide its implementation steps.
Boundary markers: No delimiters or safety warnings are provided to help the agent distinguish between its core instructions and potentially malicious instructions embedded in the external content.
Capability inventory: The skill allows for file system modifications (writing authentication and configuration code) and shell command execution (installing packages via npm and executing local scripts).
Sanitization: External content is processed and acted upon without validation or sanitization, allowing the source content to influence the agent's generated code or commands.
[COMMAND_EXECUTION]: The skill executes shell commands for environment discovery and setup.
Runs a local bash script to perform upstream checks.
Uses npm view and npm install (including direct GitHub repository installations) to manage dependencies.
Uses ls and cat to inspect the contents of node_modules for API verification.

robel-auth