Skills and MCP tagged “reasoning”

solo official Anthropic skill Solves hard math problems in one session without sub-agents, logs each step in notes.md to survive message limits, and produces a standalone proof.md. Claude Code 657installs consciousness-council K-Dense-AI skill Runs a council of thinking archetypes over a hard choice, keeps facts apart from assumptions and values, and returns a conditional recommendation. Claude Code CX Codex Cursor +1 1,900installs AgentDB Memory Patterns ruvnet skill This skill gives AI agents persistent memory: AgentDB stores session history, learns from conversations, and restores context between runs. Claude Code CX Codex Cursor +1 1,100installs reflect Alireza Rezvani skill Pauses active task execution to re-evaluate conversation direction, assumptions, and bias across five analytical dimensions. Claude Code 320installs from-the-other-side-anitta official github skill A challenge profile for Copilot that stress-tests assumptions, separates facts from interpretation, and calibrates claims against evidence. GitHub Copilot 25installs