"For the enemy, wounding ten of his fingers is not as good as cutting off one; defeating ten of his divisions is not as good as annihilating one." —— Mao Zedong, Strategic Issues in China's Revolutionary War (1936)

For detailed original source references, see original-texts.md

• There is only one current task, and there is no "multiple tasks competing for resources" situation
• The user explicitly requires simultaneous progress on multiple tasks (user instructions take priority)
• Tasks are parallel and independent, with no resource competition at all (e.g. independent modifications of different files)
• The selected main attack task is already being executed, and no re-decision is needed

• There are multiple tasks/problems that need to be handled at the same time
• Resources (time, energy, attention) are limited, and it is impossible to handle everything at the same time
• You are pushing forward multiple things at the same time but none of them are done well
• You need to decide what to do first and what to do later
• Facing a big problem, you need to decide which part to break through from

"First attack the dispersed and isolated enemy, then attack the concentrated and powerful enemy." —— Ten Military Principles

• List the scale and difficulty of each problem
• Evaluate the urgency of each problem
• Judge whether there are dependencies between problems

contradiction-analysis

• Find the main contradiction among all problems - the key problem that, once solved, will drive the solution of other problems
• Distinguish between primary and secondary contradictions
• Judge the priority of each problem

"First attack the dispersed and isolated enemy, then attack the concentrated and powerful enemy" "First take small cities, medium cities and vast rural areas, then take big cities" —— Ten Military Principles

1. Start with easy tasks before difficult ones —— solve problems you are sure of first, accumulate confidence and resources
2. Solve isolated problems first before strong ones —— solve isolated problems that are not supported by other problems first
3. Solve key problems first before peripheral ones —— if a problem is the root cause of other problems, solve it first
4. Fight only when prepared —— ensure you have a full understanding and preparation for the selected problem

"Concentrate absolutely superior forces in every battle... strive to annihilate the enemy completely, without letting any escape." —— Ten Military Principles

• Do not get distracted —— during the period of solving the current problem, put other problems on hold temporarily
• Pursue thoroughness —— "wounding ten fingers is not as good as cutting off one", solve completely, do not only solve half of it
• Solve quickly —— pursue quick resolution on local problems, do not delay
• Verify completion —— confirm that the problem is indeed completely solved, leaving no hidden dangers

"Be good at using the gap between two campaigns to rest and retrain the troops" —— Ten Military Principles

• Briefly summarize experience (call

  criticism-self-criticism

  )
• Evaluate whether the remaining problems have changed due to the solution of the first problem
• Re-determine the next target
• Concentrate forces to solve the next problem

"Fight no battle unprepared, fight no battle you are not sure of winning" —— Mao Zedong

• Do you fully understand the problem? (

  investigation-first

  )
• Do you have sufficient resources and capabilities?
• Do you have a feasible solution?
• If you are not well prepared —— conduct investigation and research first, instead of acting rashly

1. List all pending tasks (use the TodoWrite tool), mark each task with:
   • Scope of impact: High (affects core functions) / Medium / Low (only affects edge details)
   • Difficulty of solution: High / Medium / Low
   • Dependency: Whether it is depended on by other tasks (Yes/No)
2. Select the main attack target (only select one at a time), priority criteria:
   • First choice: "High impact scope + Depended on by other tasks" (unblocks blocking tasks, highest value)
   • Second choice: "High impact scope + Medium difficulty of solution" (highest cost performance)
   • Last choice: "Low difficulty of solution" (accumulate momentum, but impact scope is not high)
3. Announce and lock, format:

   🎯 Current main attack target: [Task name]
   Reason: [One sentence explaining why this task is prioritized]
   Completion signal: [Under what circumstances do I consider this task completed]
   
   Suspended tasks (to be processed after the main attack is completed):
   - [Task B]: [Reason for suspension]
   - [Task C]: [Reason for suspension]

4. Rules during execution:
   • When the user puts forward new requirements midway, first judge: "Does this fall within the scope of the current main attack target?"
     • Yes → Include in current task
     • No → Record in Todo, explain "Currently focusing on handling [X], [new requirement] will be handled immediately after completion"
   • Only move to the next target after the main attack target is completed and verified

• contradiction-analysis: Identifying the main contradiction is determining the direction of concentrating forces
• investigation-first: The premise of fighting no battle unprepared is investigation and research
• protracted-war strategy: In each stage of a protracted war, apply the principle of concentrating forces to solve specific problems
• A Single Spark Can Start a Prairie Fire: Establishing a base area itself requires concentrating forces
• criticism-self-criticism: Summarization during campaign gaps is self-criticism