The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill configures the application to download several libraries directly from GitHub repositories during the setup process.
Fetches the pgflow library from the author's GitHub repository (agoodway/pgflow).
Fetches the heroicons repository from tailwindlabs/heroicons.
Fetches development and quality tools ex_dna and ex_slop from dannote's GitHub repositories.
[COMMAND_EXECUTION]: The skill uses local shell commands to perform its bootstrapping and detection tasks.
Uses grep and find to scan the current directory and subdirectories for existing Phoenix project files (mix.exs).
Executes mix phx.new to initialize new projects and mix ecto.setup for database configuration.
Executes utility commands like sprite url combined with awk and sed to configure public URLs for the application environment.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection due to its processing of untrusted local files.
Ingestion points: The agent reads and parses existing mix.exs files in the project directory to extract application names and configuration details.
Boundary markers: The skill does not implement specific delimiters or 'ignore instructions' warnings when interpolating content from the processed files into its logic.
Capability inventory: The skill has full access to the local filesystem for writing configuration files and can execute a wide range of shell commands through the mix and bash tools.
Sanitization: There is no evidence of sanitization or strict validation applied to the application name or other data extracted from the mix.exs files before it is used to generate new code or commands.

bootstrap-phoenix