3.8K ·GitHub
What is this?
Qiushi-Skill is a collection of AI Agent Skills that distills one core principle (seeking truth from facts) and nine methodology tools from classical dialectical materialism and practical philosophy, aimed at making AI agents investigate first, focus on the main contradiction, validate through practice, and keep pushing until a task is actually finished.
Qiushi-Skill is a collection of AI Agent Skills that distills one core principle (seeking truth from facts) and nine methodology tools from classical dialectical materialism and practical philosophy, aimed at making AI agents investigate first, focus on the main contradiction, validate through practice, and keep pushing until a task is actually finished.Is this a standalone agent or a set of skill files I load into my existing AI?It is a set of skill specs plugged into a host agent: each methodology skill keeps only five sections (when to use / not use, procedure, output template, discipline, handoff) so it stays small enough to inject into a session rather than running as its own program.
It is a set of skill specs plugged into a host agent: each methodology skill keeps only five sections (when to use / not use, procedure, output template, discipline, handoff) so it stays small enough to inject into a session rather than running as its own program.How do I install the skills into my coding agent?Run `npx qiushi-skill`, which interactively detects installed agent hosts and writes the skill files into the right directory (per-platform target directories are listed in docs/platforms.md); it is also published on the Claude Code Marketplace.
Run `npx qiushi-skill`, which interactively detects installed agent hosts and writes the skill files into the right directory (per-platform target directories are listed in docs/platforms.md); it is also published on the Claude Code Marketplace.Do I need Node.js to install this?The recommended `npx qiushi-skill` installer needs Node.js >=18.17; if you don't have Node, you can skip it entirely and just copy the skills/ folders straight into your host's skills directory.
The recommended `npx qiushi-skill` installer needs Node.js >=18.17; if you don't have Node, you can skip it entirely and just copy the skills/ folders straight into your host's skills directory.Will every method fire on every request?No - the host AI judges per-scenario whether a given methodology tool is worth invoking; tasks that are straightforward execution don't trigger any downstream skill, and if the host already has an equivalent workflow, the host's own process takes precedence.
No - the host AI judges per-scenario whether a given methodology tool is worth invoking; tasks that are straightforward execution don't trigger any downstream skill, and if the host already has an equivalent workflow, the host's own process takes precedence.What are the nine methodology tools included?Contradiction analysis, practice-based epistemology, investigation and research, the mass line, criticism and self-criticism, protracted strategy, concentration of forces, igniting-prairie-fires (incremental base-building), and holistic coordination - each sourced from a specific classic text like 'On Contradiction' or 'On Protracted War'.
Contradiction analysis, practice-based epistemology, investigation and research, the mass line, criticism and self-criticism, protracted strategy, concentration of forces, igniting-prairie-fires (incremental base-building), and holistic coordination - each sourced from a specific classic text like 'On Contradiction' or 'On Protracted War'.Does the project include dispatchable subagents?Yes - agents/investigator.md is a read-only researcher that produces a fact / inference / unknown report, and agents/self-critic.md reviews finished work with fresh eyes, judging only the artifact and not the author's explanation.
Yes - agents/investigator.md is a read-only researcher that produces a fact / inference / unknown report, and agents/self-critic.md reviews finished work with fresh eyes, judging only the artifact and not the author's explanation.Is this project a political statement?The author explicitly frames it as methodology research, not politics or propaganda: all quotes are cited from public works purely for practical method extraction, with no political stance or reinterpretation intended.
The author explicitly frames it as methodology research, not politics or propaganda: all quotes are cited from public works purely for practical method extraction, with no political stance or reinterpretation intended.How is this different from giving the AI a historical persona?The author rejects persona simulation entirely, arguing LLMs shouldn't impersonate historical figures; the skills only package proven methods as reusable cognitive tools, not character role-play.
The author rejects persona simulation entirely, arguing LLMs shouldn't impersonate historical figures; the skills only package proven methods as reusable cognitive tools, not character role-play.What are similar projects?oil-ui, Claude-Code-Game-Studios, lieflat-gongwen, awesome-copilot, open-pstack, skills, superpowers-zh, pstack-claude
oil-ui, Claude-Code-Game-Studios, lieflat-gongwen, awesome-copilot, open-pstack, skills, superpowers-zh, pstack-claudeWhich topics is it in?methodology, subagents, agent-skill, claude-skills, workflow, prompt-engineering, skill, skills, china, agents, ai-agent, agent-skills
methodology, subagents, agent-skill, claude-skills, workflow, prompt-engineering, skill, skills, china, agents, ai-agent, agent-skills