@irwan_andesta: #fvpシ ngk ni bos😅😅yang laen cit hana sagai..asai nya lewat ciduk saja🤫🤫

@Irwan_94
@Irwan_94
Open In TikTok:
Region: ID
Monday 07 February 2022 19:05:30 GMT
220
24
0
1

Music

Download

Comments

There are no more comments for this video.
To see more videos from user @irwan_andesta, please go to the Tikwm homepage.

Other Videos

Most Claude Code /memory. Only the first 200 lines of MEMORY.md load. 3. Rules — .claude/rules/*.md with a paths: field. They only enter context when Claude touches a matching file. 🔵 CAPABILITY — what it can reach 4. Skills — a SKILL.md workflow that loads only when the task calls for it. The rest stay out of your context. 5. Subagents — .claude/agents/*.md. A second, separate context window with its own budget. 6. MCP — your docs, tickets and databases, plugged in through one protocol. 🟣 CONTROL — what you actually allow 7. Hooks — shell commands on lifecycle events. They run whatever Claude decides. 8. Permissions — allow, ask, deny. settings.json, four scopes. 9. Plugins — all of the above bundled. One install, whole setup. 👀 The one nobody says: CLAUDE.md persuades. Hooks enforce. CLAUDE.md is delivered as context, not enforced config — so there's no guarantee it's followed. If something MUST happen before every edit or commit, a PreToolUse hook is the only thing that actually stops the action. Every "why didn't it follow my instructions" post is really a hook that was never written. 📸 Screenshot the grid. Save it before your next setup. Follow @hackproduct — we turn scary AI-engineering concepts into things you can ship. ⚡ . . #claudecode #anthropic #claude #AIcoding #AIengineering " width="135" height="240">
Most Claude Code "cheat sheets" list slash commands that don't exist. ⚡ I checked. /plan, /code, /test, /refactor, /deploy — not real commands. And even if they were, slash commands are the shallowest layer of the tool. Here's what actually changes how Claude Code works for you 👇 🟢 MEMORY — what it knows before you type 1. CLAUDE.md — the rules you write. Loaded every session. Under 200 lines beats 800. 2. Auto memory — notes Claude writes for itself in ~/.claude/projects//memory. Only the first 200 lines of MEMORY.md load. 3. Rules — .claude/rules/*.md with a paths: field. They only enter context when Claude touches a matching file. 🔵 CAPABILITY — what it can reach 4. Skills — a SKILL.md workflow that loads only when the task calls for it. The rest stay out of your context. 5. Subagents — .claude/agents/*.md. A second, separate context window with its own budget. 6. MCP — your docs, tickets and databases, plugged in through one protocol. 🟣 CONTROL — what you actually allow 7. Hooks — shell commands on lifecycle events. They run whatever Claude decides. 8. Permissions — allow, ask, deny. settings.json, four scopes. 9. Plugins — all of the above bundled. One install, whole setup. 👀 The one nobody says: CLAUDE.md persuades. Hooks enforce. CLAUDE.md is delivered as context, not enforced config — so there's no guarantee it's followed. If something MUST happen before every edit or commit, a PreToolUse hook is the only thing that actually stops the action. Every "why didn't it follow my instructions" post is really a hook that was never written. 📸 Screenshot the grid. Save it before your next setup. Follow @hackproduct — we turn scary AI-engineering concepts into things you can ship. ⚡ . . #claudecode #anthropic #claude #AIcoding #AIengineering

About