The Agent Skills Directory

[SAFE] (SAFE): No malicious code, network activity, or sensitive data access was detected. The skill is entirely composed of documentation and instructional guidelines.
Prompt Injection (LOW): The 'Expert Panel Review' section in SKILL.md uses mandatory constraints (e.g., 'Only return when 90+ average achieved') to steer the model's reasoning process. While this is a form of behavioral instruction, it does not target safety filters or attempt instruction overrides.
Indirect Prompt Injection (LOW): The skill is designed to process untrusted data such as user-provided web pages and funnel descriptions. Ingestion points: User-provided marketing context in SKILL.md. Boundary markers: Not explicitly defined. Capability inventory: No executable capabilities (no subprocesses, no file-writes, no network ops). Sanitization: None detected. Risk: Low, as the potential impact is limited to the quality of generated text recommendations.

cro