The Agent Skills Directory

[SAFE]: The skill consists entirely of Markdown guidelines and a JSON evaluation file. It does not contain any executable scripts, command-line operations, or network requests.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill instructions include reading local context files such as .agents/product-marketing.md. These are standard project-level configuration files and do not involve sensitive system credentials or unauthorized data access.
[REMOTE_CODE_EXECUTION]: No patterns for remote code execution or external package installations (e.g., npm, pip, curl | bash) were found.
[OBFUSCATION]: A manual and automated scan for Base64, zero-width characters, homoglyphs, and hidden text patterns yielded no results.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied marketing briefs, it lacks any dangerous capabilities (such as file system writes or network access) that could be exploited via malicious input. The risk of indirect injection is negligible as the output is restricted to text generation.

copywriting