The Agent Skills Directory

[SAFE]: The skill operates as a legitimate search aggregator, directing traffic only to established and well-known services like DuckDuckGo, Microsoft Bing, and Tavily.
[SAFE]: Credential management is handled securely; the skill retrieves API keys from environment variables or a local JSON file, which prevents the exposure of sensitive secrets within the code itself.
[SAFE]: The WebContentFetcher class in multi_search.py performs essential sanitization by removing <script> and <style> tags from fetched HTML content, which helps protect the agent from indirect prompt injection or malicious scripts.
[SAFE]: Local file operations are limited to the skill's own directory for caching network status and managing API quotas, with no evidence of unauthorized access to sensitive system paths.

multi-search