The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates TypeScript code and immediately executes it using npx playwright test <file>. This is a core functionality intended to verify that the generated test passes in a headless environment before completion.
[EXTERNAL_DOWNLOADS]: The skill depends on the @playwright/test Node.js package for its testing framework and execution environment.
[PROMPT_INJECTION]: The skill processes potentially untrusted scenario data to generate executable code, which presents an indirect prompt injection surface.
Ingestion points: Reads verified scenarios from tools like flow-runner, network-assertion, or scenario-mapper (SKILL.md).
Boundary markers: None; the skill does not explicitly use delimiters or instructions to ignore malicious content embedded within the source scenarios.
Capability inventory: The skill has the ability to write files to the tests/ directory and execute shell commands via npx.
Sanitization: The instructions do not describe sanitization or validation of the input scenario steps before they are converted into executable TypeScript code.

e2e-codegen