Yao Crux Skill logo

Yao Crux Skill

CommunityPopular
yaojingang
yao-crux-skill

Use when a user is stuck in a complex real-world situation with many conflicts, limited resources, unclear priorities, or asks to identify the principal contradiction, secondary contradictions, dominant aspect, breakthrough actions, monitoring thresholds, probability projection, and Markdown/Word/HTML/PDF reports. Triggers on 矛盾论, 主要矛盾, 次要矛盾, 主次矛盾, 卡住了, 优先级混乱, 找关键问题, 资源有限, 冲突很多. Do not use for philosophy-only summaries, Mao text commentary with no live case, generic brainstorming with no diagnosis, or final licensed medical, legal, investment, or psychological-crisis advice.

Overview

Publisheryaojingang
Repositoryyao-open-skills
Skill nameyao-crux-skill
Stars
1.3K
Forks
149
Bundled files
51
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 51 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by yaojingang on GitHub. Read the source before you install it.

Installation

Install the Yao Crux Skill AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/yaojingang/yao-open-skills.git /tmp/yao-open-skills
mkdir -p .claude/skills
cp -r /tmp/yao-open-skills/skills/yao-crux-skill .claude/skills/yao-crux-skill
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Yao Crux Skill in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Yao Crux Skill on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Yao Crux Skill is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Yao Crux Skill

Use this skill to first clarify the user's current situation, then turn it into a principal-contradiction diagnosis, a secondary-contradiction watchlist, a focused action plan, and a four-format report bundle.

Use This Skill For

  • ask concise follow-up questions before judging a user's current situation
  • assess whether the user's current situation is clear enough for diagnosis before naming the principal contradiction
  • produce a confirmable current-state snapshot with goal, facts, stage, resources, constraints, stakeholders, and repeated patterns
  • identify goals, stage, constraints, facts, stakeholders, resources, and repeated conflict patterns
  • rewrite messy symptoms as candidate contradictions in the form 力量A 与 力量B 的冲突
  • run a first-principles layer scan so the principal contradiction can be an upstream, not-yet-visible constraint rather than the loudest symptom
  • use the three plain principal-contradiction checks: 决定性, 牵引性, and 阶段性
  • separate internal changeable structures from external hard conditions, then judge how external conditions act through internal structure
  • use Bayesian-style evidence updates and Ockham-razor parsimony as supporting checks when candidate contradictions are close
  • score principal and secondary contradictions by goal impact, causal leverage, stage urgency, resource constraint, changeability, spillover risk, and evidence strength
  • identify the principal aspect inside the principal contradiction
  • recommend 1-3 breakthrough actions that explicitly weaken or reverse the principal aspect, monitoring thresholds, review conditions, and probability projections after action
  • make resource and attention allocation decisive when the principal contradiction is clear, typically concentrating 50%-70% of high-leverage resources on the main line
  • include visual reasoning modules: analysis flow, iceberg model, contradiction decision matrix, resource allocation, and stage transition map
  • write user-facing reports that explicitly use 主要矛盾 and 次要矛盾, while translating them with practical labels such as 最关键的卡点 and 先不主攻,但要盯住
  • export synchronized markdown, html, docx, and pdf reports using the Kami-style layout rules in this package

Do Not Route Here

  • generic explanation of 矛盾论 with no concrete user case
  • historical, ideological, or textual commentary where no diagnosis or action report is needed
  • broad brainstorming that does not require a principal/secondary contradiction conclusion
  • final professional medical, legal, investment, safety-critical, or psychological-crisis advice

Default Workflow

  1. Read references/intake-and-questioning.md; extract the user's target, stage, constraints, resources, facts, stakeholders, and previous attempts.
  2. Run the current-state gate from references/intake-and-questioning.md or scripts/current_state_clarity.py; score clarity before diagnosis.
  3. If clarity is insufficient, do not name a principal contradiction. Ask no more than 3 decisive questions and stop.
  4. If clarity is workable, produce a provisional current-state snapshot, ask the missing questions, and label any contradiction read as provisional.
  5. If clarity is clear, restate the current-state snapshot and ask the user to confirm or correct it before deep diagnosis.
  6. Run references/safety-and-boundaries.md; hard constraints remain guardrails even when they are not the principal contradiction.
  7. Read references/theory-anchors.md; use 决定性, 牵引性, and 阶段性 as the plain-language front gate before detailed scoring.
  8. Use the first-principles gate in references/contradiction-model.md: separate visible symptoms from the hidden root variable that can explain several symptoms together.
  9. Separate 内部可改变结构, 外部硬条件, and 外因通过内因起作用; do not mistake an external condition for the principal contradiction when an internal response path is available.
  10. Build 3-7 candidate contradictions, including at least one upstream candidate if the visible symptoms share a common cause.
  11. Score candidates with the weighted model; use Bayesian-style updates to ask which explanation fits the evidence best.
  12. Apply Ockham razor only as a supporting tie-breaker: prefer the explanation that covers more symptoms with fewer extra assumptions.
  13. If the top contradiction is below 3.5/5 or the evidence is too weak, return a provisional read and ask for the missing evidence instead of over-claiming.
  14. Identify the principal contradiction, its principal aspect, secondary contradictions, defer reasons, monitor triggers, and review conditions.
  15. Tie each breakthrough action to the principal aspect: say whether it weakens the currently dominant side, strengthens the weaker side, or changes their relationship.
  16. Reallocate resources aggressively when evidence is sufficient: protect the main line, cap secondary contradictions, and keep only minimal stop-loss capacity for monitored risks.
  17. Build a canonical report input JSON using templates/crux-report.schema.json, including visual and dynamic-stage fields when available.
  18. Run scripts/generate_report_bundle.py input_file.json output_dir for synchronized Markdown, HTML, Word, PDF, and .report.json artifacts.
  19. Check the bundle with scripts/verify_report_bundle.py output_dir before claiming completion.
  20. For HTML/PDF delivery, perform a visual pass on the conclusion, current-state tables, iceberg, decision matrix, and stage-transition sections. Reject the output if the content container width drifts away from the top navigation width, or if there is horizontal overflow, clipped legends, overlapping labels, empty chart space, or decorative imagery that does not explain the analysis.

Output Contract

  • Lead with the conclusion: current principal contradiction, principal aspect, confidence, and first action.
  • Before the conclusion, make sure the current-state snapshot is explicit enough; if not, lead with follow-up questions instead of a diagnosis.
  • Default report headings must integrate the concepts and the practical translation: 主要矛盾(最关键的卡点), 次要矛盾(先不主攻,但要盯住), 主要矛盾判断过程.
  • The report must show the first-principles reasoning: 看得见的问题, 上升一层, 看不见的根部变量, and why the visible issue is not necessarily the principal contradiction.
  • The report must explain the principal contradiction through 决定性, 牵引性, and 阶段性 before falling back to scores.
  • The report must separate 内部可改变结构 from 外部硬条件, then explain how external conditions act through internal structure.
  • The report must include the visual reasoning sections 一张图看懂:从表象到主要矛盾, 时间、精力、资源应该怎么重新分配, and 主要矛盾什么时候会转移.
  • HTML/PDF reports should render four inline SVG charts plus a photo-based iceberg module. The iceberg must use a real above-water/below-water iceberg image, not a flat glacier, abstract triangle, or decorative ice background. When the image is wide enough, place it as a full-width visual anchor, put 水面上 and 水面下 explanation cards below it, and make 当前主要矛盾 a full-width long module. Markdown and DOCX can use readable text/table fallbacks.
  • Mark major facts and numbers as observed, estimated, or assumed.
  • Keep the analysis dynamic: every principal contradiction conclusion needs a reversal condition and a review point.
  • Treat secondary contradictions as monitored risks, not ignored problems.
  • Keep action advice narrow: 1-3 actions with owner, deadline, resource, metric, and expected effect.
  • Each action should name the principal aspect it changes.
  • When diagnosis confidence is sufficient, resource allocation should be deliberately tilted rather than balanced: usually 50%-70% to the principal contradiction, 10%-25% to secondary-risk containment, and 10%-20% to monitoring/evidence.
  • Include probability projection as a decision-support estimate, not a promise: baseline probability, expected action uplift, risk drag, scenario probabilities, and sensitivity notes.
  • Reports default to Simplified Chinese and must stay synchronized across Markdown, HTML, Word, and PDF.
  • The HTML/PDF layout should follow references/kami-layout-guide.md: warm paper background, ink-blue accent, restrained hierarchy, compact tables, readable chart labels, no text overlap, and print-safe A4 rules.

Reference Map

  • references/intake-and-questioning.md: follow-up strategy and input contract
  • references/theory-anchors.md: theory-to-rule anchors for principal contradiction, internal/external causes, principal aspect, and aggressive resource allocation
  • references/contradiction-model.md: scoring model, principal aspect, secondary contradiction, and probability projection rules
  • references/report-contract.md: report sections and language standard
  • references/report-export-pipeline.md: four-format generation workflow and verification
  • references/kami-layout-guide.md: Kami-derived layout rules for the report bundle
  • references/safety-and-boundaries.md: high-risk and professional-boundary gates
  • reference-materials/README.md: downloadable background research and design materials for readers
  • templates/crux-report.schema.json: canonical input/report shape
  • scripts/generate_report_bundle.py: JSON to Markdown, HTML, DOCX, PDF, and report JSON
  • scripts/current_state_clarity.py: current-state clarity gate and follow-up question selector
  • scripts/verify_report_bundle.py: generated artifact checks

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Yao Crux Skill AI skill do?

Use when a user is stuck in a complex real-world situation with many conflicts, limited resources, unclear priorities, or asks to identify the principal contradiction, secondary contradictions, dominant aspect, breakthrough actions, monitoring thresholds, probability projection, and Markdown/Word/HTML/PDF reports. Triggers on 矛盾论, 主要矛盾, 次要矛盾, 主次矛盾, 卡住了, 优先级混乱, 找关键问题, 资源有限, 冲突很多. Do not use for philosophy-only summaries, Mao text commentary with no live case, generic brainstorming with no diagnosis, or final licensed medical, legal, investment, or psychological-crisis advice.

Why use Yao Crux Skill on TypingMind?

Because you install it once and use it with any model. Yao Crux Skill is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Yao Crux Skill in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/yaojingang/yao-open-skills/tree/main/skills/yao-crux-skill. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Yao Crux Skill?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Yao Crux Skill?

As many as you like. As long as a model supports skills, you can use Yao Crux Skill with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Yao Crux Skill AI skill free?

Yes. It is published on GitHub by yaojingang under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇