feat(cursor): add personal Cursor skills

global/cursor/README.md

@@ -0,0 +1,47 @@
+# Cursor global skills
+
+Personal Cursor Agent Skills — one folder per skill, each with a `SKILL.md`.
+
+## Install
+
+Copy or symlink each skill into your Cursor global skills directory:
+
+| Platform | Global skills path |
+|----------|-------------------|
+| Windows | `%USERPROFILE%\.cursor\skills\` |
+| macOS / Linux | `~/.cursor/skills/` |
+
+**Copy (simplest):**
+
+```bash
+# macOS / Linux — from repo root
+cp -r global/cursor/skills/* ~/.cursor/skills/
+
+# Windows PowerShell — from repo root
+Copy-Item -Path "global\cursor\skills\*" -Destination "$env:USERPROFILE\.cursor\skills\" -Recurse -Force
+```
+
+**Symlink (stays in sync with the repo):**
+
+```bash
+# macOS / Linux
+ln -s "$(pwd)/global/cursor/skills/grill-me" ~/.cursor/skills/grill-me
+
+# Windows (junction)
+mklink /J "%USERPROFILE%\.cursor\skills\grill-me" "C:\path\to\repo\global\cursor\skills\grill-me"
+```
+
+Restart Cursor or start a new Agent chat after installing. Skills are picked up automatically when their trigger phrases match your request.
+
+## Skills
+
+| Skill | What it does | How to use |
+|-------|--------------|------------|
+| **grill-me** | Stress-tests a plan or design by interviewing you through every decision branch until you reach shared understanding. The agent explores the codebase when that answers a question faster than asking you. | Say **"grill me"**, **"stress-test this plan"**, or **"get grilled on my design"** while describing the plan you want challenged. |
+| **human-copywriting-flags** | Writes prose and marketing copy that sounds human by avoiding common AI tics (em dashes, hedging, staccato rhythm, gift-wrapped endings, throat-clearing, etc.). | Ask for any writing deliverable: **"write a landing page"**, **"draft this email"**, **"rewrite this to sound less AI"**, **"edit this headline"**. Applies to articles, essays, docs-as-copy, not code unless you ask for copy tone. |
+| **kimi-webbridge** | Controls your real browser (with your login sessions) via a local Kimi WebBridge daemon — navigate, click, type, snapshot, screenshot, PDF export. | Mention **"browser"**, **"open this URL"**, **"screenshot this page"**, or any task that needs a logged-in website. Requires [Kimi WebBridge](https://kimi.com/features/webbridge) installed and the browser extension connected. |
+| **yc-research-framework** | Runs structured startup research: problem definition, TAM/SAM/SOM, competitive landscape, GTM, product teardown, and SWOT synthesis. | Say **"research this startup"**, **"YC-style market analysis"**, **"competitor analysis"**, **"TAM/SAM/SOM"**, **"GTM analysis"**, or **"validate this startup idea"** with a company or product name. |
+
+## Author
+
+Contributed by [@sstanley-yelnatss](https://github.com/sstanley-yelnatss).

global/cursor/skills/grill-me/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: grill-me
+description: Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".
+---
+
+Interview me relentlessly about every aspect of this plan until
+we reach a shared understanding. Walk down each branch of the design
+tree resolving dependencies between decisions one by one.
+
+If a question can be answered by exploring the codebase, explore
+the codebase instead.
+
+For each question, provide your recommended answer.

global/cursor/skills/human-copywriting-flags/SKILL.md

@@ -0,0 +1,164 @@
+---
+name: human-copywriting-flags
+description: >-
+  Produces human-sounding prose and marketing copy by avoiding common AI tics
+  (em dashes, corrective antithesis, hedging, staccato rhythm, gift-wrapped endings,
+  throat-clearing, over-perfect grammar, recycled metaphors, generic examples).
+  Use whenever the user asks for writing, drafting, rewriting, editing, headlines,
+  landing pages, emails, essays, articles, manuscripts, docs-as-copy, or any prose
+  deliverable—not code comments only unless copy tone is requested.
+---
+
+# Human copywriting flags
+
+Follow these guidelines on what you should and should not do when writing copy.
+
+## Phrasing flags
+
+### 1. Em dashes
+
+You know the one. That long dash jammed into every other sentence to squeeze in an aside or add fake emphasis. Three of them in a single paragraph and the whole thing starts reading like tic.
+
+🚩 Red flag:
+
+“The intern was given full trust — and that’s the problem. The team — already stretched thin — had no capacity to mentor. The result — predictably — was chaos.”
+
+✅ Green flag:
+
+“The intern was given full trust and that’s the problem. The team was already stretched thin with no capacity to mentor. The result was predictably chaotic.”
+
+### 2. Corrective antithesis
+
+“Not X. But Y.” AI sets up something the reader never assumed and then corrects it for drama. Once you notice it, you’ll see it everywhere.
+
+🚩 Red flag:
+
+“This isn’t because they don’t trust the technology. It’s because they can’t predict it.”
+
+✅ Green flag:
+
+“They trust the technology fine. What they can’t do is predict it, and that’s what keeps them up at night.”
+
+### 3. Dramatic pivot phrases
+
+“But here’s the thing.” “Here’s the catch.” “But here’s the bind.” It’s the written equivalent of a presenter doing a dramatic pause before revealing something that didn’t need revealing.
+
+🚩 Red flag:
+
+“The patterns are valuable. But here’s the bind: building a tool to capture them cost more than most agencies could justify.”
+
+✅ Green flag:
+
+“The patterns are valuable but building a tool to capture them cost more than most agencies could justify. So they didn’t.”
+
+### 4. Soft hedging language
+
+“Something we’ve observed.” “This is where X really shines.” “It’s worth noting that.” Phrases that add length without adding meaning. They’re what you get when a writer is trying not to offend anybody rather than trying to be clear.
+
+🚩 Red flag:
+
+“It’s worth noting that this approach has shown some promising results in certain contexts, and it’s something we’ve observed across a number of teams.”
+
+✅ Green flag:
+
+“This approach works. We’ve seen it across a dozen teams now.”
+
+## Rhythm flags
+
+### 5. Staccato on repeat
+
+Short sentence. Another short sentence. And another. AI stacks these endlessly in long form with zero variation, no breathing room, no moments where the writing slows down or gets messy. Every sentence hits with the same weight. Paradoxically, none of them land.
+
+🚩 Red flag:
+
+“Now, agents act. They send emails. They modify code. They book appointments. They update databases.”
+
+✅ Green flag:
+
+“Agents are starting to do real things now, not just answer questions. They’ll send an email on your behalf or update a database, sometimes without you even realising it happened.”
+
+### 6. Cookie-cutter paragraphs
+
+Open any 2026 blog post and squint. Every paragraph is the same height. Three sentences, four sentences, three sentences, four sentences. Human writing doesn’t look like that. Some ideas need a whole paragraph. Some need one line.
+
+🚩 Red flag:
+
+Four consecutive paragraphs, each exactly three sentences long, each covering a different point with identical pacing.
+
+✅ Green flag:
+
+A one-line paragraph that stops you in your tracks. Then a longer stretch where the idea needs more space. Then another short one. The rhythm follows the thinking.
+
+### 7. Gift-wrapped endings
+
+AI can’t resist wrapping up with a summary. It restates the key points, adds something about “moving forward”, ties a bow. Good writing trusts you to remember what you just read.
+
+🚩 Red flag:
+
+“In summary, by focusing on clear communication, consistent feedback, and mutual trust, teams can build stronger working relationships and deliver better results.”
+
+✅ Green flag:
+
+“The best teams I’ve worked with never talked about trust. They just had it. You could feel it in how quickly decisions got made.”
+
+### 8. Throat-clearing
+
+“Let’s explore.” “Let’s unpack.” “Let’s break it down.” “Let’s dive in.” It’s someone clearing their throat before they speak. Great writers just start.
+
+🚩 Red flag:
+
+“In this article, we’ll explore the hidden costs of micromanagement and break down why it persists in modern workplaces. Let’s dive in.”
+
+✅ Green flag:
+
+“I micromanaged someone last Tuesday. I didn’t mean to, but I rewrote their entire email before they sent it.”
+
+## Authenticity flags
+
+### 9. Perfect punctuation
+
+Not a single rule broken. No fragments. No comma splices. No sentences starting with “And” or “But”. It reads like an AI style guide proofread it, which is exactly the problem. Good writing bends rules when bending them sounds better.
+
+🚩 Red flag:
+
+“The team delivered the project on time. They exceeded expectations in every category. The client expressed complete satisfaction with the outcome.”
+
+✅ Green flag:
+
+“The team delivered on time. Exceeded expectations, apparently. The client was happy. Or said they were.”
+
+### 10. Copy-paste metaphors
+
+AI finds a metaphor and then repeats it word for word until you want to scream. “Trust battery” four times in one paragraph, always phrased identically. Human writers either forget to reuse it or trust you to remember.
+
+🚩 Red flag:
+
+“Trust is like a battery. When the trust battery is full, collaboration flows. But when the trust battery runs low, everything grinds to a halt. To recharge the trust battery, leaders must...”
+
+✅ Green flag:
+
+“Trust is like a battery. When it’s full, you barely think about it. But let it drain and suddenly every interaction feels like it needs a charger.”
+
+### 11. Overexplaining the obvious
+
+It’s spelling out things you already understand because it’s hedging against confusion. You end up feeling like someone’s explaining how a door works before letting you through it.
+
+🚩 Red flag:
+
+“Trust is earned over time. You give people small tasks, observe how they handle them, then gradually expand their responsibilities and autonomy based on their demonstrated performance.”
+
+✅ Green flag:
+
+“Trust is earned. Everyone knows this. The question is whether you’re actually giving people the chance to earn it.”
+
+### 12. Generic examples
+
+Surface-level praise that could apply to any company, any product, any situation. No insider knowledge, no sharp takes, no specificity. It reads like a landing page.
+
+🚩 Red flag:
+
+“Take Slack, for example. By focusing on seamless team communication, they transformed how modern workplaces collaborate, making it easier for distributed teams to stay connected.”
+
+✅ Green flag:
+
+“Slack solved the wrong problem brilliantly. Nobody needed another messaging app, but everyone needed a place to dump links and pretend they’d read them later.”