The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill operates by fetching and analyzing untrusted external data using the web_fetch tool to perform SEO audits, SERP analysis, and AEO visibility testing. While this is the intended purpose of the skill, it lacks explicit boundary markers or sanitization logic to prevent the agent from following instructions potentially embedded in the fetched HTML or third-party content. This ingestion surface, combined with tools for command execution (bash) and file modification (write_to_file, replace_file_content), creates a vulnerability to multi-step prompt injection attacks.\n
Ingestion points: External website content and SERP data retrieved via web_fetch across all audit and optimization modes (referenced in SKILL.md, references/aeo.md, references/technical.md, and references/onpage.md).\n
Boundary markers: Absent; the instructions do not provide specific delimiters or 'ignore' directives for processed web data.\n
Capability inventory: The skill environment includes bash, view_file, write_to_file, replace_file_content, and grep_search.\n
Sanitization: Absent; there is no documented procedure for filtering or escaping external content before the agent processes it or acts upon its analysis.

seo