The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill is purely instructional and utilizes local project context (.agents/product-marketing-context.md) for brand alignment, which is a standard and secure practice for marketing agents.
[PROMPT_INJECTION]: The skill processes user-supplied text for editing. Ingestion points: user-provided copy for review. Boundary markers: absent. Capability inventory: none (no subprocess, network, or file system write access). Sanitization: absent. Because the skill lacks any exploitable tools or executable code, the indirect injection surface represents no functional security risk.

copy-editing