The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes keyword data and search snippets from external search engines, which represents an attack surface for indirect prompt injection.\n
Ingestion points: SKILL.md describes expanding seeds via WebSearch and DataForSEO (Step 1 and Step 2).\n
Boundary markers: No specific boundary markers or instructions to the agent to ignore embedded commands are provided when processing search results.\n
Capability inventory: The skill possesses file writing capabilities (cluster-plan.json), execution of local plugin scripts, and interaction with the claude-blog skill for content creation.\n
Sanitization: The skill uses a url_safety mode for fetching external URLs via its internal rendering script.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute local scripts for specialized tasks like cost estimation and page rendering.\n
Evidence: SKILL.md contains instructions to run dataforseo_costs.py and render_page.py located in the plugin's script directory.\n- [DYNAMIC_EXECUTION]: The skill generates an interactive visualization by injecting external data into an HTML template at runtime.\n
Evidence: Step 6 of the planning workflow involves reading templates/cluster-map.html and replacing a JavaScript variable with a JSON object constructed from search data. This results in the creation of a local file (cluster-map.html) for the user to open.\n- [EXTERNAL_DOWNLOADS]: The skill documentation suggests installing a related skill from the author's GitHub repository.\n
Evidence: The skill documentation links to github.com/AgriciDaniel/claude-blog if the skill is missing. As this resource belongs to the skill's author, it represents a vendor-owned resource.

seo-cluster