The Agent Skills Directory

[SAFE]: The skill is designed purely for text manipulation and copy editing. It lacks capabilities for network access, file system modification (beyond reading context), or command execution.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of user-provided marketing copy. While this presents an attack surface for prompt injection, the risk is negligible because the skill does not possess any dangerous tools or output channels that could be exploited by such an injection.
Ingestion points: User-provided copy requested in SKILL.md.
Boundary markers: None present.
Capability inventory: Text refinement and generation. No file-writing, network, or execution capabilities.
Sanitization: None present.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill instructions include loading a local configuration file .agents/product-marketing-context.md. This is a legitimate use of internal context to define brand voice and does not involve exfiltrating data to external sources.

copy-editing