Codified systems for shipping growth — auditable by humans, executable by agents.
Every growth tactic has a half-life.
What worked last quarter breaks this quarter. Outreach moved from email to Twitter DM to GitHub issues. Content moved from blog posts to threads to AI-native formats. Seed users moved from waitlists to competitor poaching. Every shift, the surface tactics get easier — and the underlying craft gets harder.
X-Studio captures the systems that survive the decay. Each folder is something tested in a real campaign. Each skill is written so an AI agent can run it without hand-holding, and a growth operator can read it before pressing run.
This isn't a course. It isn't a blog. It's a bench — the place where the work happens, opened up.
| Folder | What it is |
|---|---|
skills/ |
Atomic, agent-executable units of growth work — drop into Cursor / Claude Code / Codex |
playbooks/ |
Narrative SOPs — how multiple skills compose into a real campaign |
research/ |
Field reports and ecosystem snapshots — operator interpretation, not journalism |
templates/ |
Copy-paste assets — cold emails, landing pages, issue forms |
case-studies/ |
Real campaigns — methods, data, what worked, what didn't |
Twelve atomic skills currently shipped. Each has been used in a live campaign before being written down.
| # | Skill | What it does |
|---|---|---|
| 1 | stargazer-email-scraper |
Extract public emails of competitor-repo stargazers — 15–17% hit rate |
| 2 | competitive-intel-mining |
Mine competitor communities (GitHub / Discord / Reddit) for high-intent users |
| 3 | competitive-analysis |
Deep-dive a competitor / category — growth signals, community pulse, GitHub bursts, SEO + strategic implications |
| 4 | competitor-deep-research-sop |
SOP v1.3.1 — user-pain first research: M1×M2 matrix, 14 dimensions, 12-layer evidence + sub-channels, S1–S3 seed signals, Tier-1 list, monitoring; ships with full 930-line methodology reference |
| 5 | seed-user-pool-stratification |
Score and segment seed users by ROI before spending a single message |
| 6 | personalized-cold-outreach |
1-to-1 personalized outreach at batch scale — first three sentences must be true only for that recipient |
| 7 | landing-to-issue-funnel |
Turn a cold email into a pre-filled GitHub Issue capture loop |
| 8 | content-repurpose-engine |
One insight → bilingual threads, long-form, video hooks |
| 9 | growth-experiment-template |
Structured experiment doc — hypothesis, metric, decision rule, retro |
| 10 | reddit-account-warmup |
Warm a fresh Reddit account into a stable, low-risk identity in 6 weeks |
| 11 | community-content-bot |
Bilingual daily content push to Feishu + Discord on GitHub Actions — 60-min ship, sub-$10/mo cost |
| 12 | ai-agent-sdd |
Write a professional Software Design Document (SDD) for an AI agent before vibe-coding — MVP (12 sec) and Full (23 sec) templates, AI-specific sections (LLM strategy / agent workflow / eval framework / failure modes) |
Roadmap teaser — linkedin-content-warmup, x-account-warmup, discord-community-seeding, outbound-followup-sequencer (each will land when shipped at least once).
git clone https://github.com/Celina-create/X-Studio.git ~/x-studio
# Mount everything
ln -s ~/x-studio/skills ~/.claude/skills/x-studio
# Or a single skill
ln -s ~/x-studio/skills/competitive-intel-mining ~/.cursor/skills/competitive-intel-miningInvoke by name in chat:
"Use
competitive-intel-miningon this competitor's GitHub issues."
"Runreddit-account-warmupand tell me which week I'm in."
Nothing here is copy-protected. Steal what's useful. A star is appreciated, not required.
Four principles that gate what gets published here.
- Has shipped at least once. No theoretical playbooks. If it hasn't been run on a real campaign, it doesn't make it in.
- ROI over volume. Every skill optimizes for precision. The bottom 50% of any "list" is usually noise — we cut it before any message is sent.
- Agents are operators. Every skill is structured so a coding agent can execute the full chain. Hand-off points are explicit. State is explicit.
- Method over mystique. Public method, public reasoning, public limits. Only client / user data is redacted.
| Today | |
|---|---|
| Skills shipped | 11 |
| Playbooks | 6 (1 acquisition + 1 community + 4-piece open-source DevRel system) |
| Research reports | 1 (Feishu × OpenClaw ecosystem snapshot) |
| Case studies | 1 (Bilingual content bot post-mortem) |
| License | MIT |
This repository updates as work happens — small commits, frequent additions. Watch the repo for new skills as they land.
Built by Celina (@Celina-create) — a growth operator who treats outreach, content, and seeding as engineerable systems rather than vibes.
For collaboration, custom skill requests, or notes exchange:
MIT © 2026 Celina. Use, fork, remix, attribute when convenient.
Growth is a craft. This is the bench.