The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for running shell commands using the bun CLI (e.g., bun bd test test/bundler/bundler_banner.test.ts). These commands are standard for executing test suites in the Bun development environment.
[DYNAMIC_EXECUTION]: The testing framework described (itBundled) is designed to take JavaScript and TypeScript code strings, bundle them, and execute them to verify output. This involves runtime compilation and execution of code snippets provided in the files configuration object.
[INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting code snippets into a test runner. If an agent uses these instructions to process untrusted external data (such as code from a GitHub issue or pull request) without sanitization, it could execute malicious code embedded within those snippets.
Ingestion points: Code strings within the files and runtimeFiles objects in SKILL.md examples.
Boundary markers: None explicitly defined in the provided templates to separate untrusted code from test logic.
Capability inventory: Execution of shell commands via bun bd test and file system operations within the test framework.
Sanitization: No specific sanitization or validation of the input code strings is mentioned in the guide.

writing-bundler-tests