The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates a temporary JavaScript file (screenshot-script.mjs) using a template and executes it with node. This involves dynamic string interpolation of user-supplied application and login URLs without explicit sanitization, which could lead to code injection or unexpected execution flow.
[COMMAND_EXECUTION]: The skill executes shell commands to check for Playwright dependencies (npm ls), create the screenshots directory (mkdir), and run the generated script (node).
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local codebase data to identify features and URLs. | Ingestion points: README.md, CHANGELOG.md, package.json, and various routing configuration files (e.g., config/routes.rb, app/ routes) as specified in Step 3. | Boundary markers: Absent. The skill instructions do not specify any delimiters or warnings to ignore instructions embedded in the ingested data. | Capability inventory: Subprocess calls for package checking, file system modification (mkdir), and execution of a generated Node.js script that launches a browser (SKILL.md). | Sanitization: Absent. Content from the codebase is directly used to populate the list of features, paths, and CSS selectors for the screenshot automation.

screenshots