The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection. It ingests untrusted data from external websites (HTML content via render_page.py) and protocol configurations (JSON via ucp_check.py) without explicit boundary markers or instructions to ignore embedded commands. This could allow an attacker to influence the agent's behavior by placing malicious instructions on the analyzed pages.
[COMMAND_EXECUTION]: The skill relies on several external scripts (render_page.py, parse_html.py, dataforseo_merchant.py, ucp_check.py) to perform its core functions. While these are part of the skill's expected environment, they represent an execution surface for processing externally supplied data.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to fetch content from arbitrary user-provided URLs and makes requests to the DataForSEO Merchant API.
[DATA_EXFILTRATION]: The ucp_check.py script includes a --probe-endpoints flag that performs network probes (HEAD requests) against URLs discovered within a site's /.well-known/ucp file. This creates a potential Server-Side Request Forgery (SSRF) surface if the target site's profile contains internal or malicious URLs, although the documentation notes that SSRF-blocked endpoints are reported.

seo-ecommerce