The Agent Skills Directory

[OBFUSCATION]: The skill includes Base64 encoded strings and hex escapes within its documentation, notably in the XSS and JWT sections of SKILL.md. These are presented as inert examples for security testing (e.g., <svg onload=eval(atob('YWxlcnQoMSk='))>) and do not obfuscate the skill's operational logic.\n- [DATA_EXFILTRATION]: SKILL.md documents sensitive local file paths (such as /etc/passwd and ~/.aws/credentials) and cloud metadata endpoints as reference payloads for SSRF and LFI testing. These references are consistent with the skill's purpose as an offensive security arsenal and do not involve active data exfiltration by the skill itself.\n- [EXTERNAL_DOWNLOADS]: REFERENCES.md contains links to well-known security community repositories on GitHub. These references are documented neutrally and serve as additional research resources for penetration testing workflows.

security-arsenal