The Agent Skills Directory

[SAFE]: The skill implements strong Server-Side Request Forgery (SSRF) mitigations in scripts/ghost.py. It resolves the target hostname and explicitly verifies that all resolved IP addresses are globally routable, preventing the agent from being used to scan or attack internal network services (e.g., localhost, 169.254.169.254).
[SAFE]: Secure transport is strictly enforced by requiring the GHOST_SITE_URL to use the HTTPS protocol and validating that the URL contains no embedded credentials, queries, or fragments that could lead to information leakage or path confusion.
[SAFE]: The skill employs a 'dry-run' architecture for data-modifying operations. The Python script will only perform POST or PUT requests to the Ghost API if the --confirm flag is explicitly provided, serving as a critical safety gate against accidental or unauthorized writes.
[SAFE]: The implementation uses a custom PinnedHTTPSConnection to ensure that network requests are sent specifically to the IP addresses validated during the SSRF check, while still performing standard TLS certificate verification against the original hostname.
[SAFE]: The skill exclusively uses the Python standard library, eliminating risks associated with malicious or vulnerable third-party dependencies.

ghost