qiushi-skill

Overview of qiushi-skill

What it is

Qiushi-Skill is an open-source collection of AI Agent Skills. It extracts one general principle and nine methodological tools from classical materialist dialectics and practical philosophy. The project aims to systematically equip an AI's reasoning process, not with slogans, but with an actionable methodology set. Each method is supported by direct citations from classic texts, which are available in the original-texts.md files within each skill directory. The project is not about politics or propaganda; it is about methodology for analyzing and solving problems. It is also not a personality distillation; the author states that large models currently cannot and should not be used to simulate human personalities.

Key features

• Provides a total principle of Seeking Truth from Facts: judgments must follow evidence, distinguish facts/inferences/unknowns, only count verified work as done, and diagnose obstacles before giving up.
• Includes nine methodological tools organized in three layers: philosophical foundation (Contradiction Analysis, Practice-Cognition Epistemology), working methods (Investigation, Mass Line, Criticism & Self-Criticism), and strategy & tactics (Protracted Strategy, Concentrate Forces, Spark Prairie Fire, Overall Planning).
• Each skill is a concise, actionable module with five sections: trigger conditions, operating procedures, output templates, discipline rules, and handoff to the next skill.
• Supports multiple AI agent platforms including Claude Code, Cursor, Codex, OpenCode, OpenClaw, Hermes, and Nanobot.
• Includes subagents (investigator.md and self-critic.md) that can be dispatched in independent contexts.
• Provides a CLI tool (npx qiushi-skill) for easy installation and validation across platforms.

Use cases

• Analyzing complex problems by identifying the main contradiction.
• Validating and iterating on solutions through the practice-cognition cycle.
• Gathering information before making decisions using the investigation method.
• Integrating feedback and validating plans with the mass line approach.
• Reviewing work and improving quality through criticism and self-criticism.
• Planning long-term, complex tasks with a protracted strategy.
• Making priority decisions and focusing resources by concentrating forces.
• Developing strategies from scratch using the spark prairie fire method.
• Balancing multiple objectives and making trade-offs with overall planning.

Getting started

bash npx qiushi-skill For Claude Code Marketplace: text /plugin marketplace add HughYau/qiushi-skill /plugin install qiushi-skill@qiushi-skill

License and pricing

The project is licensed under MIT. It is open-source and free to use.

Based on the project's public README and repository metadata.