The Agent Skills Directory

[COMMAND_EXECUTION]: The skill workflow involves executing a multi-line JavaScript snippet via node -e to process test results (SKILL.md, references/extraction-workflow.md). This command reads from local files and performs parsing and data extraction.
[REMOTE_CODE_EXECUTION]: In assets/analytics-common-step.ts, the skill uses require() with dynamic paths constructed from component names (require(./${component}/${component}.json)). This creates a vulnerability where malicious input in a tracking specification could potentially cause path traversal or the loading of unintended local modules.
[EXTERNAL_DOWNLOADS]: The skill installs allure-js-commons, allure-playwright, and playwright-bdd from the npm registry for reporting and BDD support. It also references fetching configuration data from Confluence's REST API.
[PROMPT_INJECTION]: The skill exhibits an attack surface for Indirect Prompt Injection by processing data from external documentation URLs to drive automated test generation.
Ingestion points: Retrieves tracking specifications from wiki pages or documentation URLs provided by the user.
Boundary markers: No delimiters or safety instructions are used to distinguish external data from the agent's core instructions.
Capability inventory: The skill performs file system writes, subprocess execution (npx), dynamic code loading (require), and browser-level script injection (page.evaluate).
Sanitization: External data is not validated or sanitized before being incorporated into the test code and execution workflow.

datalayer-analytics-playwright