The Agent Skills Directory

[SAFE]: The skill fetches data from well-known security research services including HaveIBeenPwned (haveibeenpwned.com), DeHashed (api.dehashed.com), and IntelX (2.intelx.io). These are established platforms for verifying credential leaks and are considered safe sources for the skill's intended purpose.
[SAFE]: The password check implementation in check_password uses the k-Anonymity model, which is the recommended privacy-preserving method for the Pwned Passwords API. Only the first five characters of a SHA-1 hash are sent to the service.
[SAFE]: The skill's primary function involves sending email addresses and domains to external APIs for breach verification. This is the intended behavior and uses standard HTTP requests to known providers.
[PROMPT_INJECTION]: The skill has an indirect prompt injection attack surface because it processes and displays data retrieved from external API sources. 1. Ingestion points: Untrusted data enters the agent context through API responses in hibp_request, check_dehashed, and check_intelx functions. 2. Boundary markers: The skill does not use delimiters or instructions to the agent to ignore potential commands embedded in the API data. 3. Capability inventory: The skill has capabilities for network operations via the requests library and file system writing via json.dump (specifically writing breach_report.json in bulk_check_emails). 4. Sanitization: No validation or sanitization is performed on the data returned from the external APIs before it is printed or saved.

breach-data