π―Skills33
An agent built around a ReAct chat loop with file-backed memory and geofenced subagents. Each turn reassembles the system prompt from persistent state files (persona, self-model, user model, discourse agreements) and runs tool-augmented reasoning with post-turn reflection.
A skill from the Tons of Skills marketplace, a large open-source collection of over 200 Claude Code plugins and skills that can be installed and managed through the ccpi command-line tool.
Part of Cognitive Workbench, an agent system built around a ReAct chat loop with file-backed memory and geofenced subagents. Each turn reassembles the system prompt from persistent state files including persona, user model, discourse agreements, and active concerns.
A ReAct-loop chat agent with file-backed persistent memory, a companion user model, and geofenced subagents. Each turn assembles a system prompt from persona, self-model, discourse agreements, and recalled memories, then runs reflection to update state between turns.
A ReAct-loop agent with file-backed persistent memory, companion modeling, and geofenced subagents. Each turn assembles a system prompt from persona, self-model, discourse agreements, recalled memories, and active concerns, then runs a tool loop for text processing, search, and reflection.
A ReAct-loop agent with file-backed persistent memory, reflection between turns, and a geofenced subagent architecture that reassembles its system prompt from persona, self-model, and discourse state files each turn.
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
