The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from site-config.json to generate React components and configuration files.
Ingestion points: Data is read from site-config.json (Step 1) and $ARGUMENTS (Step 2).
Boundary markers: None identified. The instructions do not include specific delimiters or warnings to ignore instructions embedded within the configuration data.
Capability inventory: The skill uses Bash, Write, and Edit tools to create files and execute shell commands, including npm install and npm run build (Step 11).
Sanitization: There is no mention of escaping or validating the content of the site-config.json before it is interpolated into .jsx, .html, or .json files. If the configuration contains malicious instructions or escape characters, it could lead to code injection or unintended command execution during the build process.
[COMMAND_EXECUTION]: The skill executes npm install and npm run build (Step 11) within a directory controlled by the configuration. While standard for web development, executing these commands on scaffolded code that incorporates untrusted input carries a risk of executing malicious scripts defined in a generated package.json or through dependency confusion.
[REMOTE_CODE_EXECUTION]: The skill performs an npm install (Step 11), which fetches and executes code from the npm registry. While the base dependencies (react, react-dom) are standard, the process relies on the integrity of the configuration data and the local environment.

website-build