The Agent Skills Directory

[SAFE]: The skill demonstrates secure credential handling practices. In scripts/lib.mjs, the attachOriginAuth function specifically scopes authentication headers to a single origin via Playwright route filters. This prevents sensitive tokens (like SITE_TOKEN or DA_TOKEN) from being leaked to third-party domains during site scraping or analysis.
[SAFE]: The identified network operations and domain references (such as aem.live, hlx.page, and da.live) are official Adobe and Edge Delivery Services infrastructure. These are consistent with the skill's primary purpose of migrating sites to these platforms.
[SAFE]: The use of dynamic execution through page.evaluate and browser automation via Playwright is a standard and necessary implementation for a tool designed to analyze client-side rendered content and dynamic behaviors.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by crawling websites. While this creates a surface for indirect prompt injection, the risk is mitigated because the tool performs structured data extraction (e.g., finding form fields, script tags, and API paths) rather than using the LLM to interpret arbitrary prose from the source site. The severity is internally assessed as LOW and does not escalate the final verdict.

dynamics