The Agent Skills Directory

[SAFE]: The script scripts/fetch_page.py incorporates robust SSRF mitigations by validating that URLs use allowed protocols and ensuring that resolved IP addresses are global and not part of private or reserved ranges.
[INDIRECT_PROMPT_INJECTION]: The skill instructions and shared contracts explicitly define external data as untrusted, instructing the agent to ignore any embedded commands within ingested content, which effectively manages the risk of indirect prompt injection.
[SAFE]: All included scripts, such as those for report generation and health score calculation, perform deterministic tasks on local data without employing dangerous functions like eval() or exec().
[SAFE]: The suggestion to install the standard reportlab library for PDF generation is handled transparently and does not involve automated or unverifiable remote code execution.

seo-hreflang