學習路徑
s01 到 s20：漸進式 Agent Harness 設計
層次圖例
The Agent LoopOne Loop Is All You Need
“The smallest useful agent is a loop that calls the model, runs tools, and feeds results back.”
瞭解更多Tool UseAdd a Tool, Add Just One Line
“The loop stays stable while capabilities register into a dispatch table.”
瞭解更多PermissionCheck Permissions Before Execution
“Dangerous actions need a harness decision point before the shell runs.”
瞭解更多HooksHang on the Loop, Don't Write into It
“Cross-cutting behavior belongs around the loop, not tangled inside it.”
瞭解更多TodoWriteAn Agent Without a Plan Drifts Off Course
“Explicit plans keep long-running work visible and correctable.”
瞭解更多SubagentBreak Large Tasks into Small Ones with Clean Context
“Subagents give each subtask a clean message history while preserving the main thread.”
瞭解更多Skill LoadingLoad Only When Needed
“Inject specialized knowledge only when the task actually needs it.”
瞭解更多Context CompactContext Will Fill Up
“Compression keeps the conversation usable when the context window gets crowded.”
瞭解更多MemoryKeep a Layer That Doesn't Lose Details
“Some facts should survive summarization and future sessions.”
瞭解更多System PromptAssembled at Runtime, Never Hardcoded
“The system prompt is a generated product of policy, tools, skills, and context.”
瞭解更多Error RecoveryErrors Are the Start of a Retry
“A robust harness classifies failures and decides what kind of retry is worthwhile.”
瞭解更多Task SystemBreak Big Goals into Small Tasks
“A task graph turns vague goals into ordered, observable work.”
瞭解更多Background TasksSlow Operations Go to the Background
“The agent can keep reasoning while slow work completes elsewhere.”
瞭解更多Cron SchedulerProducing Work on a Schedule
“Recurring work should be created by the harness, not remembered by the model.”
瞭解更多Agent TeamsOne Agent Isn't Enough, Form a Team
“Persistent teammates let work continue in parallel without stuffing every thought into one context.”
瞭解更多Team ProtocolsTeammates Need Agreements
“Multi-agent systems need explicit message contracts, not vibes.”
瞭解更多Autonomous AgentsCheck the Board, Claim the Task
“Teammates become useful when they can discover and claim work themselves.”
瞭解更多Worktree IsolationSeparate Directories, No Conflicts
“Parallel agents need isolated filesystems as much as isolated conversations.”
瞭解更多MCP ToolsExternal Tools, Standard Protocol
“External services can become agent tools through a standard discovery and call protocol.”
瞭解更多Comprehensive AgentAll Mechanisms, One Loop
“The final harness is still one loop, now surrounded by the systems that make it production-shaped.”
瞭解更多CLAUDE.mdPut Your Work Rules in Writing
“AI doesn't disobey — it guesses when you leave blanks. A written manual removes the guessing.”
瞭解更多VerificationDon't Trust 'Done' — Verify It
“Fluent text is not the same as correct output. Define 'done' before the task starts.”
瞭解更多MemoryNew Chat, Same Context
“Memory lives in your files, not in AI's head. A three-part handoff note bridges any conversation gap.”
瞭解更多HooksWhen Rules Aren't Enough, Install a Gate
“CLAUDE.md is a guide. A hook is a guard. If 5% failure is unacceptable, enforce it.”
瞭解更多Layered GovernanceToo Many Rules? Sort Them into Drawers
“AI ignores the second half of a 200-line rule file. Split into global / task / personal layers.”
瞭解更多IntegrationAssemble a System That Actually Works at the Office
“Start with a minimum viable harness — manual, checklist, handoff note — and add pieces only when needed.”
瞭解更多