The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes large volumes of untrusted data from user-provided URLs, creating a surface for indirect prompt injection where malicious site content could attempt to manipulate the agent's audit results or behavior.
Ingestion points: External website content and metadata fetched via user-provided URLs (found in SKILL.md).
Boundary markers: No explicit delimiters or boundary markers for external content were identified in the instruction set.
Capability inventory: Execution of localized Python scripts for environment management, file system writes to the '.seo-cache' directory, and network access to various SEO-related service APIs.
Sanitization: The instructions do not detail specific techniques for sanitizing HTML or structured data parsed from external sources.
[COMMAND_EXECUTION]: The skill is designed to execute specific internal scripts located within its repository to perform tasks such as authentication checks and drift history monitoring (e.g., 'scripts/google_auth.py' and 'scripts/drift_history.py').
[EXTERNAL_DOWNLOADS]: The skill relies on well-known services and APIs for its core functionality, such as Google PageSpeed Insights, Moz, and Bing Webmaster Tools, which are recognized industry standards for SEO data.

seo