The Agent Skills Directory

[SAFE]: The skill implements a robust multi-agent architecture with built-in quality and security gates. A dedicated 'Counter-Spy' agent (Skeptic) validates every deliverable for provenance and accuracy, significantly mitigating risks associated with processing untrusted external data during research phases.
[SAFE]: Input sanitization is performed on user-provided arguments (such as company names) to derive 'slugs' (alphanumeric and hyphens only). This prevents path traversal or command/prompt injection when these strings are used to generate file paths and agent identifiers.
[SAFE]: The skill includes explicit 'Shared Principles' that prohibit agents from including secrets, credentials, or personally identifiable information (PII) in prompts, messages, or checkpoint files.

profile-competitor