{"id":50288139,"url":"https://github.com/HughYau/qiushi-skill","last_synced_at":"2026-06-13T20:01:00.504Z","repository":{"id":346782056,"uuid":"1191520000","full_name":"HughYau/qiushi-skill","owner":"HughYau","description":"求是Skill——从经典唯物辩证法与实践哲学中提炼出一条总原则和九大方法论工具武装AI大脑。Qiushi-Skill: Build agents that investigate first, focus on the main contradiction, validate in practice, and keep pushing until the work is actually done.","archived":false,"fork":false,"pushed_at":"2026-04-29T06:38:48.000Z","size":3814,"stargazers_count":2892,"open_issues_count":18,"forks_count":221,"subscribers_count":9,"default_branch":"main","last_synced_at":"2026-04-29T08:31:50.263Z","etag":null,"topics":["ai-agents","methodology","skills","workflow"],"latest_commit_sha":null,"homepage":"http://hughyau.com/qiushi-skill/","language":"JavaScript","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/HughYau.png","metadata":{"files":{"readme":"README.en.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-25T10:24:01.000Z","updated_at":"2026-04-29T08:29:31.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/HughYau/qiushi-skill","commit_stats":null,"previous_names":["hughyau/qiushi-skill"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/HughYau/qiushi-skill","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HughYau%2Fqiushi-skill","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HughYau%2Fqiushi-skill/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HughYau%2Fqiushi-skill/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HughYau%2Fqiushi-skill/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/HughYau","download_url":"https://codeload.github.com/HughYau/qiushi-skill/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HughYau%2Fqiushi-skill/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34298252,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-13T02:00:06.617Z","response_time":62,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-agents","methodology","skills","workflow"],"created_at":"2026-05-28T04:00:27.319Z","updated_at":"2026-06-13T20:01:00.499Z","avatar_url":"https://github.com/HughYau.png","language":"JavaScript","funding_links":[],"categories":["哲学家 \u0026 思想家 Philosophers","Public Figures and Methodological Perspectives"],"sub_categories":["AI研究者"],"readme":"\u003cp align=\"center\"\u003e\n  \u003cimg src=\"https://raw.githubusercontent.com/HughYau/qiushi-skill/main/assets/logo_main.png\" width=\"400\"/\u003e\n\u003c/p\u003e\n\n# Qiushi Skill: 'Seeking Truth' Skills for AI Agents\n\n\u003cp align=\"center\"\u003e\n  \u003cstrong\u003eLanguages\u003c/strong\u003e:\n  \u003ca href=\"./README.md\"\u003e简体中文\u003c/a\u003e |\n  \u003ca href=\"./README.en.md\"\u003eEnglish\u003c/a\u003e\n\u003c/p\u003e\n\n\u003e Build agents that investigate first, focus on the main contradiction, validate in practice, and keep pushing until the work is actually done.\n\n`qiushi-skill` is a collection of reusable methodology skills, commands, and session-start hooks. It turns a set of problem-solving methods into installable assets for Claude Code, Cursor, Codex, OpenCode, OpenClaw, Hermes Agent, and other prompt-driven hosts.\n\n## Why This Exists\n\nMany agents can generate text, but they still fail at disciplined work:\n\n- they answer before collecting facts\n- they scatter effort across too many tasks\n- they stop at the first blocker\n- they skip verification and call that \"done\"\n\n`qiushi-skill` is meant to correct that with one core principle and nine concrete methods:\n\n- `seeking truth from facts` as the governing principle\n- contradiction analysis\n- practice -\u003e cognition -\u003e practice iteration\n- investigation first\n- mass line feedback synthesis\n- criticism and self-criticism\n- protracted strategy\n- concentrate forces\n- spark a prairie fire\n- overall planning\n\n\n\u003e Support author's other repo!: **[AcademicForge](https://github.com/HughYau/AcademicForge)**: **One Forge, All Skills.** An interactive, cross-platform tool to generate simple, one-click installation commands for multiple AI Agent Skills. \n\n## Install\n\n### Preferred: `npx qiushi-skill`\n\n```bash\nnpx qiushi-skill\n```\n\nNon-interactive examples:\n\n```bash\nnpx qiushi-skill install --target claude-code --scope user\nnpx qiushi-skill install --target claude-code,cursor --scope project\nnpx qiushi-skill install --target codex,opencode,openclaw,hermes,nanobot --scope user\nnpx qiushi-skill install --target all --scope user\nnpx qiushi-skill validate\nnpx qiushi-skill uninstall --target claude-code --scope user\n```\n\nThe CLI is zero-dependency and does four things:\n\n- installs standard plugin bundles for Claude Code and Cursor\n- copies skills/commands into the directories scanned by Codex, OpenCode, OpenClaw, Hermes, and nanobot\n- writes `.qiushi-skill-install.json` manifests so uninstall only removes CLI-managed files\n- validates the current checkout or published bundle with one Node-based code path\n- keeps `install`, `validate`, and `uninstall` under one stable entrypoint\n\n### Claude Code Official Marketplace\n\nThis repository now ships `.claude-plugin/marketplace.json`, so Claude Code can discover it directly from GitHub:\n\n```text\n/plugin marketplace add HughYau/qiushi-skill\n/plugin install qiushi-skill@qiushi-skill\n```\n\n### Source Install\n\n```bash\ngit clone https://github.com/HughYau/qiushi-skill\ncd qiushi-skill\nclaude --plugin-dir .\n```\n\n### Platform Notes\n\n- Claude Code: use the GitHub marketplace flow above, or `npx qiushi-skill install --target claude-code`.\n- Cursor: use `npx qiushi-skill install --target cursor`, or copy the bundle to your configured plugins directory.\n- Codex: use `npx qiushi-skill install --target codex`, or see [`.codex/INSTALL.md`](.codex/INSTALL.md).\n- OpenCode: use `npx qiushi-skill install --target opencode`, or see [`.opencode/INSTALL.md`](.opencode/INSTALL.md).\n- OpenClaw: use `npx qiushi-skill install --target openclaw`, or see [`.openclaw/INSTALL.md`](.openclaw/INSTALL.md).\n- Hermes Agent: use `npx qiushi-skill install --target hermes`, or see [`.hermes/INSTALL.md`](.hermes/INSTALL.md).\n- nanobot: use `npx qiushi-skill install --target nanobot`, or see [`.nanobot/INSTALL.md`](.nanobot/INSTALL.md).\n\n## Validate\n\nPreferred:\n\n```bash\nnpx qiushi-skill validate\n```\n\nIf you are working from a source checkout, legacy fallback scripts are also available:\n\n```bash\nbash tests/validate.sh\n# or on Windows\npowershell -NoLogo -NoProfile -ExecutionPolicy Bypass -File tests/validate.ps1\n```\n\nValidation checks:\n\n- JSON validity, including `.claude-plugin/marketplace.json`\n- version consistency between `package.json` and marketplace metadata\n- presence of CLI files, hooks, commands, packaged platform guides, and the isolated marketplace bundle\n- frontmatter completeness for skills, commands, and agents\n- local Markdown link integrity\n\n## How to Use It\n\nThe session-start entry skill is `skills/arming-thought/SKILL.md`. Its job is intentionally narrow:\n\n1. enforce `seeking truth from facts`\n2. choose downstream skills only when they clearly help\n3. avoid loading every method by default\n\nManual command entrypoints live in `commands/*.md` for hosts that support Markdown slash commands.\n\n## Platforms\n\n- Claude Code: official marketplace + automatic SessionStart injection + commands\n- Cursor: bundle metadata + commands + standard CLI install path\n- OpenClaw: can map Claude/Cursor/Codex bundles into native OpenClaw plugins\n- Hermes Agent: native `skills` directory and CLI/toolset workflow\n- Codex: document-based setup\n- OpenCode: document-based setup\n- Generic hosts: reuse `skills/` and `commands/`\n\n## Project Layout\n\n```text\nqiushi-skill/\n├── .claude-plugin/\n│   ├── marketplace.json\n│   └── plugin.json\n├── .cursor-plugin/plugin.json\n├── .codex/INSTALL.md\n├── .opencode/INSTALL.md\n├── .openclaw/INSTALL.md\n├── .hermes/INSTALL.md\n├── bin/\n├── skills/\n├── commands/\n├── hooks/\n├── agents/\n├── docs/\n├── CHANGELOG.md\n├── README.md\n└── README.en.md\n```\n\n## License\n\nMIT\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHughYau%2Fqiushi-skill","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FHughYau%2Fqiushi-skill","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHughYau%2Fqiushi-skill/lists"}