{"id":51304,"date":"2026-03-21T09:38:52","date_gmt":"2026-03-21T01:38:52","guid":{"rendered":"https:\/\/www.1ai.net\/?p=51304"},"modified":"2026-03-19T16:20:17","modified_gmt":"2026-03-19T08:20:17","slug":"%e5%ae%89%e8%a3%85openclaw%e5%b0%8f%e7%99%bd%e6%95%99%e7%a8%8b%ef%bc%8copenclaw%e9%ab%98%e7%ba%a7%e7%af%87%e6%95%99%e5%ad%a6","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/51304.html","title":{"rendered":"Installation of openclaw little white course, OpenClaw advanced course"},"content":{"rendered":"<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-51307\" title=\"621b421aj00tc4z40005ad000iw007kp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/03\/621b421aj00tc4z40005ad000iw007kp.jpg\" alt=\"621b421aj00tc4z40005ad000iw007kp\" width=\"680\" height=\"272\" \/><\/p>\n<p>Mid-level solves three problems: I can't remember, I can't find it, I can't break it<\/p>\n<p>but you might find that-- angent's getting smart<\/p>\n<p>It's like you trained an intern with good memory, fast search, and no break. But you still have to tell him everything<\/p>\n<p>This is the problem that seniors are going to solve: making interns regular employees<\/p>\n<p>What do the regulars need<\/p>\n<p>Skill -- let it understand your business<\/p>\n<p>SOP(Hook) -- makes it follow the rules<\/p>\n<p>MCP - TO ENABLE IT TO USE COMPANY TOOLS<\/p>\n<p>Independently responsible for the project (autonomous mission) - let it work on its own<\/p>\n<p>And because it's complicated, we're talking about the core concept in each direction, plus a killer case<\/p>\n<p><strong>Skill \u2014 pre-service training for staff<\/strong><\/p>\n<p>you tell angent every day about the format of the daily newspaper: first the work done, then the important decision, and then the plan for tomorrow<\/p>\n<p>i told you 50 times, change the session<\/p>\n<p>no, it's not. it's not smart. it's everything you teach<\/p>\n<p>It's a question of nature: what is taught in dialogue is temporary<\/p>\n<p><strong>Skill, what is it<\/strong><\/p>\n<p>One sentence: Skill = a training manual for angent, which will take effect permanently<\/p>\n<p>It's essentially a SKILL.md file under skills\/catalogue. Scanning this directory on startup, identifying triggers, matching the content of SKILL.md into context<\/p>\n<p><strong>and\u00a0<\/strong><strong>A<\/strong><strong>GENTS.md<\/strong><strong>What<\/strong><strong>Different<\/strong><\/p>\n<p>AGENTS.md is a universal working code -- how to write, how to search, how to record mistakes. Every time a session is loaded, whatever you do<\/p>\n<p>Skill is a specific business process and knowledge -- what's your daily format, what's your code code and how your search tools work. Load only when required, match by trigger word<\/p>\n<p>AGENTS.md, working attitude, Skyll, business knowledge<\/p>\n<p>Or Skill, whatever<\/p>\n<p><strong>Murder cases: a daily newspaper, Skyll<\/strong><\/p>\n<p>When there's no Skyll, every time you write a daily, you have to say a big part:<\/p>\n<p>Write for me today's daily newspaper, in the form of a list of the main tasks accomplished today, followed by important decisions, problems encountered and solutions, and finally tomorrow's plan<\/p>\n<p>read today's memory file extract<\/p>\n<p>Fifty words of instruction. Say it every day<\/p>\n<p>Create a file after Skill ~\/.opencllaw\/skills\/ Daily-report\/SKILL.md<\/p>\n<p>&#8212;<\/p>\n<p>name: daily-report<\/p>\n<p>description: &gt;<\/p>\n<p>Generally works reports by reproducing memory files and semmarizing<\/p>\n<p>i'm sorry, acts.<\/p>\n<p>Triggers: \u201cDaily report\u201d, \u201cDaily newspaper\u201d, \u201cWhat has been done today\u201d, \u201cSummary of Work\u201d.<\/p>\n<p>&#8212;<\/p>\n<p># Daily Report Skill<\/p>\n<p>When tried:<\/p>\n<p>Read today's memory file<\/p>\n<p>2. Subject key actors, experts, and issues<\/p>\n<p>Format as:<\/p>\n<p>\u2013 Work accomplished<\/p>\n<p>\u2013 Important decision-making<\/p>\n<p>\u2013 Problems encountered<\/p>\n<p>\u2013 Tomorrow<br \/>\n&#8212;<\/p>\n<p>Name: Daily coverage<\/p>\n<p>Description:<\/p>\n<p>Daily work reports are generated by reading memory documents and summarizing activities\u3002<\/p>\n<p>Trigger words: \u201cDaily report\u201d, \u201cDaily newspaper\u201d, \u201cwhat has been done today\u201d, \u201cSummary of work\u201d\u3002<\/p>\n<p>-# DAILY REPORTING SKILLS<\/p>\n<p>When triggered:<\/p>\n<p>1. Read memory files for the day (memory\/YYYY-MM-DD.md)<\/p>\n<p>2. Extracting key activities, decision-making and issues<\/p>\n<p>The format is:<\/p>\n<p>\u2013 Work accomplished<\/p>\n<p>\u2013 Important decision-making<\/p>\n<p>\u2013 Problems encountered<\/p>\n<p>\u2013 Tomorrow<\/p>\n<p>From now on, you just have to say three words: a daily<\/p>\n<p>angent matches the trigger word, automatically loads SKILL.md, knows what to read, in what format, how to output<\/p>\n<p>fifty words to three words. and no matter how many sessions, the format is the same<\/p>\n<p>Skill doesn't exist alone, it and Hook and MCP form a complete angent enhancement system<\/p>\n<p>Let's do a preview. We'll do it in each direction<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-51308\" title=\"0b80bf2cj00tc4zvi001ld000se00fqp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/03\/0b80bf2cj00tc4zvi001ld000se00fqp.jpg\" alt=\"0b80bf2cj00tc4zvi001ld000se00fqp\" width=\"1022\" height=\"566\" \/><\/p>\n<p>Skill, whatever it is, Hook, whatever it is, MCP, whatever it is<\/p>\n<p>Remember that. It'll be used again and again<\/p>\n<p>Full Skill development tutorials -- from catalogue structure to trigger mechanism to pit-to-pit record -- in an advanced complete section<\/p>\n<p><strong>ii. Hook - SOP for employees<\/strong><\/p>\n<p>every time you open a new session, you have to say, \"look at today's plan and the diary.\"<\/p>\n<p>I've said it a hundred times and I've finally figured out one thing<\/p>\n<p>I'm not supposed to remember this<\/p>\n<p>I'm the boss, not the nymph. I shouldn't have reminded the secretary to turn on the notebook before every meeting<\/p>\n<p>Hook, what is it? Hook = angent reflect arc<\/p>\n<p>What's a reflection arc? Knee jump a reflection. The hammer hits the knee, the leg goes off, the signal goes through the spinal cord, not through the brain<\/p>\n<p>Hook, this is the thing -- when a given event occurs, the code runs, without going through ant, without LLM decision-making<\/p>\n<p>This and Skill are two levels:<\/p>\n<p>Skill is a manual for angent -- it may look like it doesn't look at it, it may look like it doesn't. First event, code runs, 1001 TP3T confirmed<\/p>\n<p>Key difference: Hook runs TypeScript code, not LLM decision-making<\/p>\n<p>ONE IS THE RECOMMENDATION, ONE IS THE IRON LAW. YOU SAY TO THE EMPLOYEE, \"EMAIL FIRST EVERY MORNING,\" HE PROBABLY FORGETS. YOU GIVE HIM A SOP WITH AN ALARM. HE CAN'T FORGET IT<\/p>\n<p>The structure of Hook is simple:<\/p>\n<p>you know, my-look\/<\/p>\n<p>Ideas--HOOK.md # announces what to listen to<\/p>\n<p>Ideas-handler.ts # on the trigger<\/p>\n<p>HOOK.md told the system, \"What's my name, what's my story?\"<\/p>\n<p>handler.ts, tell the system, \"what exactly is the incident doing?\"<\/p>\n<p>Still Case: A Hook Let angent read today's plan automatically<\/p>\n<p>every time you open a new session, you want angent to automatically read today's diary file -- you don't have to<\/p>\n<p>Synchronising \"%s\"<\/p>\n<p>&#8212;<\/p>\n<p>name: session-loader<\/p>\n<p>file: \"new session automatically loads today's working log\"<\/p>\n<p>metadata:<\/p>\n<p>openclaw:<\/p>\n<p>participants: [\"command:new\"]<\/p>\n<p>well, let's see<\/p>\n<p>bins: [`node']<\/p>\n<p>&#8212;<br \/>\n&#8212;<\/p>\n<p>Name: Session Loader<\/p>\n<p>description: \"new session automatically loads today 's working log\"<\/p>\n<p>Metadata:<\/p>\n<p><a href=\"https:\/\/www.1ai.net\/en\/tag\/openclaw\" title=\"[See articles with [OpenClaw] label]\" target=\"_blank\" >OpenClaw<\/a>:<\/p>\n<p>Event: [\"Order: New\"]<\/p>\n<p>request: bins: [`node']<\/p>\n<p>&#8212;<\/p>\n<p>_other organiser<\/p>\n<p>import {readFileSync, examplesSync} from \u201cfs\u201d;<\/p>\n<p>import {join} from \u201cpath\u201d;<\/p>\n<p>=async(event)=&gt;<\/p>\n<p>if (event. type! = \u201ccommand\u201d ||event.action! = \u201cnew\u201d) {<\/p>\n<p>return;<\/p>\n<p>}<\/p>\n<p>(a) Const memoryDir = join (proceed.cwd(), \u201cmemoory\u201d);<\/p>\n<p>const today = new Date().toISOString().split (\u201cT\u201d) [0];<\/p>\n<p>const diaryPath = join;<\/p>\n<p>if (existsSync (diaryPath))<\/p>\n<p>(a) Constcontent = readFileSync (diaryPath, \u201cutf-8\u201d);<\/p>\n<p>`today's work log has been loaded:<\/p>\n<p>${content} `;<\/p>\n<p>\u266a orse \u266a<\/p>\n<p>(a) events.messages.push (\u201cno work logs available today, start a new day\u201d);<\/p>\n<p>}<\/p>\n<p>};<\/p>\n<p>i don't know;<br \/>\nImport {readFileSync, externalsSync} from \"fs\";<\/p>\n<p>import {join} from 'path ';<\/p>\n<p>consthandler = step (incident) = {<\/p>\n<p>if! == sync, corrected by elderman == = \u201cnew\u201d)<\/p>\n<p>Return;<\/p>\n<p>}<\/p>\n<p>(a) Const memoryDir = join (proceed.cwd(), \u201cmemoory\u201d);<\/p>\n<p>const today = new Date().toISOString().split (\u201cT\u201d) [0];<\/p>\n<p>const diaryPath = join;<\/p>\n<p>If<\/p>\n<p>contContent = readFileSync (diaryPath, \u201cutf-8\u201d);<\/p>\n<p>('today's work log loaded:<\/p>\n<p>${content} ');<\/p>\n<p>Otherwise<\/p>\n<p>(a) events.messages.push (\u201cno work logs available today, start a new day\u201d);<\/p>\n<p>}<\/p>\n<p>};<\/p>\n<p>Export the default processor;<\/p>\n<p>effect: you open a new session, angent already knows what to do today<\/p>\n<p>You don't have to say, \"Look at today's plan.\"<\/p>\n<p>You don't need to be TypeScript. Just describe what needs -- \"do what at what time\" -- and get your angent to write for you handler.ts<\/p>\n<p>Core competency is not a writing code, it is defined clearly. This is what PM does every day<\/p>\n<p><strong>Progress scene<\/strong><\/p>\n<p>Hook can do much more than load the diary:<\/p>\n<ul>\n<li>Listen to campaign event: autosave task status before compression, autorestore after compression -- plan file mode for mid-text, make it fully automatic with Hook<\/li>\n<li>Listening write operation: Intercept dangerous document coverage -- 345 KB no more accidents<\/li>\n<li>Complementing Heartbeat: autochecking for unfinished tasks every heartbeat<\/li>\n<\/ul>\n<p>from \"you remind angent\" to \"agent knows himself\" -- that's the power of the reflection arc<\/p>\n<p>Several practical Hook Configuration Templates + Full Penetration is recorded in advanced complete pages<\/p>\n<p><strong>MCP - SYSTEM ACCESS FOR STAFF<\/strong><\/p>\n<p>you want angent to help you check the data in the database, and it says, \"i don't have that ability.\"<\/p>\n<p>You want it to run a specific directory in the file system. It says, \"I can't.\"<\/p>\n<p>there's only two things you can do by default: think and talk. nothing else<\/p>\n<p>You're gonna have to take it out<\/p>\n<p>MCP = Mode Context Protocol<\/p>\n<p>ONE SENTENCE: USB INTERFACE FOR THE AI WORLD<\/p>\n<p>Whatever external services -- databases, search engines, file systems, browsers -- support the MCP protocol and get on your agent<\/p>\n<p>YOU DON'T HAVE TO WRITE A SPECIFIC DOCKING CODE FOR EACH SERVICE. IT'S LIKE USB UNITES EXTERNAL INTERFACES -- WHETHER YOU'RE A PRINTER OR A CAMERA, THE SAME PLUG-IN<\/p>\n<p>LEARN USB, AND YOU WON'T HAVE TO RETRAIN ANY EQUIPMENT. MCP, TOO<\/p>\n<p>The structure is simple:<\/p>\n<p>Agent \u2192 mcporter \u2192 MCP Server \u2192 External service<\/p>\n<p>mcporter is the medium for communication between antent and MCP Server. OpenClaw incorporates mcporter skill, angent knows how to call it<\/p>\n<p>profiles in config\/mcporter.json - each project can access different external services<\/p>\n<p><strong>killer case: a configuration for angent to read and write files<\/strong><\/p>\n<p>The simplest MCP Server is the file system -- no API key, no local run<\/p>\n<p>write a configuration in config\/mcporter.json:<\/p>\n<p>{<\/p>\n<p>\"mcpServers\": {<\/p>\n<p>\"filesystem\":<\/p>\n<p>\"command\": \"npx\",<\/p>\n<p>\"args\": [\"-y\", \"@modelcontextprotocol\/server-filesystem\", \"\/tmp\/mcp-test\"]<\/p>\n<p>}<\/p>\n<p>}<\/p>\n<p>}<\/p>\n<p>{<\/p>\n<p>\"mcpServers\": {<\/p>\n<p>\"filesystem\":<\/p>\n<p>\"ORDER\": \"NPX\"<\/p>\n<p>\"args\": [(\"-y\", \"@modelcontextprotocol\/server-filesystem\", \"\/tmp\/mcp-test\")] }<\/p>\n<p>}<\/p>\n<p>}<\/p>\n<p>Three lines is enough. Meaning: Start a file system MCP Server, only allowed access to \/tmp\/mcp-test directory<\/p>\n<p>The last path is the safe boundary -- the limit for Server can only read and write documents under this directory. And a reason to give app permission<\/p>\n<p>Let's check it out<\/p>\n<p>mcporter list<\/p>\n<p>1 TP5T, seeing the filesystem Server and the tools it exposes, means it works<\/p>\n<p>=\/tmp\/mcp-test\/test.txt<\/p>\n<p># READ TEST FILES<\/p>\n<p>You can say in angent conversations, \"Help me read files under \/tmp\/ mcp-test\" -- angent will automatically go mcporter call<\/p>\n<p>Unable to answer MCP angent<\/p>\n<p>Agent of MCP can work with tools<\/p>\n<p>That's what system access means<\/p>\n<p><strong>Ecology<\/strong><\/p>\n<p>MCP is not just a file system. Server:<\/p>\n<ul>\n<li>File system (reading and writing local files)<\/li>\n<li>PostgreSQL \/ SQLite (database query)<\/li>\n<li>Brave Search (search engine)<\/li>\n<li>Puppeteer (engineer automation)<\/li>\n<li>Memoory<\/li>\n<\/ul>\n<p>A USB interface, unlimited outside. With a new Server, just a few lines of configuration in mcporter.json -- the process is exactly the same as the file system above<\/p>\n<p>Don't use MCP for MCP. A simple file reading and writing agent accompanies a tool, MCP suitable for a scenario that requires external API access, standardized multiple external calls<\/p>\n<p><strong>Autonomous tasks \u2014 independent responsibility for the project<\/strong><\/p>\n<p>Say, \"Do me a community trend study, Agent says, \"Okay, what's the first step?\"<\/p>\n<p>He's waiting for your command<\/p>\n<p>You said the first step was to search GitHub, and it did. Then we'll wait for your second step<\/p>\n<p>You said step two search Reddit, it search. Then I'll wait for you<\/p>\n<p>That's not what you want. You want to give it a goal. It dismantles its own steps, implements it, and decides whether or not it's done<\/p>\n<p>You say one step at a time -- this is an intern<\/p>\n<p>You do it to the target -- this is a regular employee<\/p>\n<p>Core concept: from directive to objective<\/p>\n<p>It's the biggest cognitive upgrade from mid-level to advanced<\/p>\n<p><strong>Command Drive:<\/strong>you give step, angent execution. this is the medium-level plan document model -- you create the plan manually, you list the steps, you go the steps<\/p>\n<p><strong>Target drivers:<\/strong>you give the target, angent deciphers his own mission, plans his own execution sequence, and makes his own judgment<\/p>\n<p>This will take three infrastructure, and the medium level has taught it:<\/p>\n<ul>\n<li>schedule file -- angent created and updated itself, not for you<\/li>\n<li>Heartbeat -- wake up angent every 30 minutes, let it see if there's something to do<\/li>\n<li>AGENTS.md Rules -- Tell angent how to decompose, how to execute, how to assess a target-type mission<\/li>\n<\/ul>\n<p><strong>killer case: to a target, angent broke six steps to implement<\/strong><\/p>\n<p><strong>actual case: i gave an agent a goal -- to complete a six-direction collection<\/strong><\/p>\n<p>all i'm saying is that we're going to take the six directions of the prompt, and we're going to do the research and the field<\/p>\n<p>what did he do:<\/p>\n<ol>\n<li>Read the target document and understand the specific requirements of the six directions<\/li>\n<li>created its own plan document<\/li>\n<li>Dismantling of 6 directions into 6 phases with clear outputs for each phase<\/li>\n<li>Start from the first direction - search for documents, field tests, recording of pedals<\/li>\n<li>Automatically update the progress of planned documents at each stage of completion<\/li>\n<li>Heartbeat regularly check to make sure angent doesn't forget what he's doing<\/li>\n<\/ol>\n<p>center context has been compressed twice and crossed three sessions<\/p>\n<p>but every time a new session starts, angent reads the plan document, then the last progress continues. i haven't lost any progress<\/p>\n<p>FINAL OUTPUT OF 5 MATERIAL DOCUMENTS ADDED UP TO 160 KB<\/p>\n<p>Mid-level program document + Heartbeat = Infrastructure, Advanced Autonomy = Building Autonomy on Infrastructure<\/p>\n<p>middle: you create the plan document, angent executes the plan document, advanced: angent creates the plan document itself, executes the plan<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-51309\" title=\"ec4b4de3j00tc50df0018d000iw007tp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/03\/ec4b4de3j00tc50df0018d000iw007tp.jpg\" alt=\"ec4b4de3j00tc50df0018d000iw007tp\" width=\"680\" height=\"281\" \/><\/p>\n<p>The ability to grow is now working<\/p>\n<p>not all missions are suitable for target-driven \u2014 simple things give orders faster. but when the task is so complicated that you don't know how many steps to take, let angent try<\/p>\n<p>Full autonomous task configuration + pedals are recorded in full, advanced content<\/p>\n<p>Four pieces of equipment, one sentence recalling:<\/p>\n<p>Skill, make it business<\/p>\n<p>Hook, let it follow the rules<\/p>\n<p>MCP LET IT USE TOOLS<\/p>\n<p>Autonomy makes it work on its own<\/p>\n<p>From mid- to high-level, we learned not four new concepts -- but the ability to raise them<\/p>\n<p>MEDIUM-LEVEL SEARCH STRATEGY \u2192 ADVANCED TOOL SELECTION<\/p>\n<p>mid-level plan document \u2192 advanced-level self-programming (agent created and updated)<\/p>\n<p>Mid-level Heartbetat<\/p>\n<p>The basics teach you how to fit, the medium ones teach you to work, the advanced ones teach you to work<\/p>","protected":false},"excerpt":{"rendered":"<p>Mid-level solves three problems: can't remember, can't find, can't break, but you might find -- angent's smart, and you have to tell him what to do one step at a time, like you trained an intern, with a good memory, a quick search, a line. But you've got to tell him everything. Skill -- let it understand your business, SOP -- make it do business by its own rules, system terms of reference (MCP) -- make it work on its own with the company's own tools, and because it's really complicated, we talk about the core concept in each direction, plus a case of a killer<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[8229,8392],"collection":[8384],"class_list":{"0":"post-51304","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-jiaocheng","7":"category-baike","8":"tag-openclaw","10":"collection-openclaw"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/51304","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=51304"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/51304\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=51304"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=51304"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=51304"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=51304"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}