The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing reference documentation for the Webflow Browser API. It includes a Python utility, scripts/search_references.py, for searching these local documents.
[SAFE]: The scripts/search_references.py script includes proactive security measures:
Path Validation: It uses pathlib to ensure that the search directory and the files it reads are contained within the skill's root directory, preventing directory traversal attacks.
Prompt Injection Sanitization: It contains a regular expression (_INJECTION_RE) designed to identify and remove common prompt injection patterns (e.g., 'ignore previous instructions', 'you are now') from the content of the reference files before they are displayed. This prevents instructions within the data from being treated as commands by the agent.
[SAFE]: The consent management examples (OneTrust, TrustArc) demonstrate standard web development practices for reading cookies and localStorage to synchronize user privacy preferences with the Webflow API. No data exfiltration or malicious intent was found.
[SAFE]: All external references point to the author's official GitHub repository (github.com/224-industries), which aligns with the provided vendor context.

webflow-browser-api