4/ The day Three features shipped. Type-checks and lint clean on every commit. Plus my first run at parallel agents: two of them on two tickets at once, each gated by the same standard before merge.…

This is where @cursor_ai gets interesting. Not “write me an app.” More like: 1. brutal architecture review 2. identify drift + duplication 3. create a branch 4. extract route into modules 5. run…

(3/9) 工具怎么执行：`anthropics/claude-code` README 定位是 terminal 里的 agentic coding tool，也可用于 IDE / GitHub `@claude`。 我更关注它的工程外壳：plugins、hooks、agents、MCP、git workflow、数据/安全边界。 注意：不要把它说成普通开源库。

New CTO Detected $PLUGINS(Cursor Plugins) CPC25cpSsUJekwbDF97ew23FdHAmdWmnpL6LRUSVpump 🔗 Quick Swap: Volume: $3.2K with 40 Liq: 43.34 SOL | MC: $14K X: Tg: Web:

Launching $PLUGINS with @ericzakariasson from the @cursor_ai team onboard! Eric is the dev behind cursor's plugin ecosystem and his github repo has 1400 stars you can check it out on the official…

cursor just shipped a plugin spec with an sdk and some official plugins. continual learning, team kits, agent compatibility.. feels like theyre building a whole ecosystem.

Cursor shipped a /thermo-nuclear-code-review for the TOUGHEST AI code review possible. But is it any good? Let's dive in:

new pstack principle: this was actually a cool emergent behavior i noticed while tackling a really large refactor to @stylexjs. my agent using /poteto-mode started writing codemods and custom skills…

been using /thermo-nuclear-code-quality-review from the @cursor_ai team kit and it’s awesome. it flags code that works right now but makes the codebase shittier later. catches big files, weird…

/quintet: for each feature / fix use a subagents. each subagent needs to be reviewed, tested, and verified by at least 4 subagents one of the subagent should use this skill has been v successful in…