The Agent Skills Directory

[SAFE]: The skill provides a structured template and clear rules for generating test code. It does not attempt to access sensitive files, perform unauthorized network operations, or bypass safety guidelines.
[COMMAND_EXECUTION]: The workflow requires the agent to verify that the generated TypeScript files compile without errors (Step 6, SKILL.md). This typically involves the local execution of standard development tools like the TypeScript compiler (tsc), which is routine for code-generation utilities.
[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to read and analyze the target page in the application to understand its elements (Step 4, SKILL.md). This exposes the agent to untrusted data from the application's DOM, which could theoretically contain instructions to influence code generation. 1. Ingestion points: The DOM of target web pages being analyzed. 2. Boundary markers: None specified to delimit application content from agent instructions. 3. Capability inventory: Creating and writing files to the local e2e/poms/ directory and executing compilation checks. 4. Sanitization: The skill relies on Playwright locators but does not mandate explicit sanitization of extracted text labels.

create-pom