The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions focus on defining the agent's persona as an expert conversion copywriter and providing strict writing rules to avoid 'AI tells'. There are no instructions that attempt to override safety guidelines or bypass constraints.
[DATA_EXPOSURE]: Analysis of the skill body shows it references local product marketing context files (e.g., .agents/product-marketing.md) to inform its drafting process. No sensitive file paths, environment variables, or hardcoded credentials were detected.
[OBFUSCATION]: No evidence of Base64 encoding, zero-width characters, or homoglyph attacks was found in any of the analyzed files. All URLs and text are in plain, human-readable format.
[REMOTE_CODE_EXECUTION]: The skill does not attempt to download or execute external scripts, and no package dependencies (Python/Node.js) are defined in the files.
[INDIRECT_PROMPT_INJECTION]: The skill defines a clear ingestion point for data via the product-marketing context files (e.g., in SKILL.md). However, the skill lacks dangerous capabilities like shell execution or network requests, and its primary purpose is constrained to text generation, which satisfies safety requirements. There is no evidence of unsafe interpolation of untrusted data into executable contexts.

copywriting