The Agent Skills Directory

[Prompt Injection] (LOW): The skill is susceptible to Indirect Prompt Injection (Category 8) because it processes untrusted data from external websites.
Ingestion points: scripts/seo_audit.py and various curl commands in SKILL.md fetch and display HTML metadata (titles, descriptions, H1 tags) from arbitrary URLs.
Boundary markers: Absent. The scripts output the extracted strings directly to the agent's context without using protective delimiters or 'ignore' instructions.
Capability inventory: The skill has the ability to execute shell commands, interact with a paid SEO API, and open browser windows.
Sanitization: Minimal. The script uses regex for extraction but does not sanitize the content of the tags for adversarial instructions.
[Command Execution] (SAFE): The skill utilizes local Python scripts and standard system commands like curl and grep. These operations are well-documented in the SKILL.md and are necessary for the skill's primary purpose of auditing web pages.
[External Downloads] (SAFE): Network communication is directed to the DataForSEO API and user-provided URLs for analysis. The script scripts/credential.py correctly retrieves credentials from environment variables (DATAFORSEO_LOGIN, DATAFORSEO_PASSWORD) rather than hardcoding them, adhering to security best practices.

seo-geo