The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions and metadata were analyzed for bypass attempts, role-play injections, or instructions to ignore safety guidelines. No such patterns were detected. All instructional text is focused on legitimate marketing workflows.
[DATA_EXFILTRATION]: No network operations (e.g., curl, wget, fetch) or hardcoded credentials (API keys, tokens) were found. The skill does not access sensitive system paths such as SSH keys or configuration files.
[COMMAND_EXECUTION]: The skill includes Python scripts (brand_voice_analyzer.py and seo_optimizer.py) for content analysis. These scripts utilize only standard libraries (re, json, sys) and do not execute shell commands or use dangerous functions like eval() or exec().
[REMOTE_CODE_EXECUTION]: No remote scripts are downloaded or executed. The skill contains no package installation commands (e.g., pip install) or dynamic loading of external code.
[SAFE]: The skill's codebase is transparent and aligns with its stated purpose of assisting in marketing content creation. The scripts perform read-only text analysis on local content without making external requests or writing to the filesystem.

content-creator