The Agent Skills Directory

[SAFE]: The skill consists of markdown templates and Python scripts that perform local text processing. No external network requests, hardcoded credentials, or obfuscated code were detected.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it is designed to process untrusted text files for SEO and brand voice analysis.
Ingestion points: Untrusted text data is ingested through the brand_voice_analyzer.py and seo_optimizer.py scripts via the existing_content.txt and blog_post.md files.
Boundary markers: Absent; the scripts read and process the full content of the input files provided as command-line arguments.
Capability inventory: The skill uses standard file reading and regex-based analysis. No subprocess spawning or dynamic code evaluation (eval/exec) is present in the provided scripts.
Sanitization: No specific sanitization or escaping of the input text is performed before processing, though the scripts only output statistical analysis and recommendations back to the agent.

content-creator