The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external websites and search engine results using WebFetch, which creates a vulnerability to indirect prompt injection. If a target business website or search result contains malicious instructions, they could potentially influence the agent's behavior during analysis.
Ingestion points: Scrapes content from business websites and Bing Maps results as described in the NAP verification and GBP audit workflows in SKILL.md.
Boundary markers: The instructions do not define the use of clear delimiters or instructions to the model to ignore any embedded directives in the fetched content.
Capability inventory: The skill has permissions to write analysis reports to the filesystem (MAPS-ANALYSIS-*.md), manage a local data cache (.seo-cache/), and perform network requests via platform-provided tools.
Sanitization: No sanitization, filtering, or validation logic is specified for the data retrieved via web scraping.
[DATA_EXFILTRATION]: The skill performs outbound network operations to several external domains to retrieve SEO and geographic data. These include bing.com, overpass-api.de, geoapify.com, and openstreetmap.org. While these are well-known services required for the skill's stated purpose, they are outside the standard whitelisted domains for data exfiltration analysis.

seo-maps