The Agent Skills Directory

[SAFE]: The skill defines a psychological and cognitive framework for managing conversation dynamics. It focuses on modulating the agent's tone and strategy based on perceived user states.
[COMMAND_EXECUTION]: The skill references the use of InfraNodus MCP tools, such as optimize_text_structure, analyze_text, and generate_topical_clusters. These tools are intended for structural discourse analysis within the vendor's own ecosystem and do not involve arbitrary shell command execution.
[DATA_EXFILTRATION]: There are no indicators of unauthorized network activity or credential harvesting. Data processing is limited to analyzing conversation text through configured MCP tools for structural insights.
[PROMPT_INJECTION]: The instructions provide guidance on behavioral shifts (e.g., being more playful or challenging assumptions) but do not contain instructions to bypass safety filters, disregard previous rules, or reveal system prompts.
[REMOTE_CODE_EXECUTION]: The skill does not perform any external downloads or execute remote scripts. It relies on internal logic and local markdown files in a references/ directory for its theoretical foundation.

cognitive-variability