The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data such as metadata, sitemaps, and JSON-LD content, which could potentially contain adversarial instructions designed to subvert agent behavior.
Ingestion points: Reads local Next.js project files including metadata components, robots.txt, and sitemap XML files.
Boundary markers: The workflow defines clear separation between analysis, preview, and modification tasks, allowing for user review between steps.
Capability inventory: The skill requires file system read/write access for analysis and fixing SEO issues, as well as CLI execution for project builds.
Sanitization: Instructions mandate checking recommendations against official Next.js documentation and current project versions to ensure technical accuracy.
[SAFE]: The configuration utility in scripts/profile_store.py includes a security check that explicitly filters out sensitive terms such as 'token', 'secret', and 'password' from being persisted to the long-term user profile.
[SAFE]: The skill uses PyYAML for configuration management, which is a standard industry dependency, and performs atomic file operations to ensure the integrity of its local storage.
[SAFE]: No remote code execution patterns, unauthorized network exfiltration, or obfuscated content were detected within the skill's scripts or instructions.

lov-better-seo