1 · /idea-forge
Spark it
Turn project context into a focused improvement backlog.
Free and open source · MIT
From rough idea to release.
A skill for every step.
52 Claude Code skills for choosing the work, building it, finding bugs, improving the experience, and shipping. Every skill is free to read, download, change and share.
Free · MIT license · Requires Claude Code
The 30-second tour
From vibe coded to shipped.
Prompting from scratch gets you a demo that breaks. The pack gives Claude Code one command for each step, and the skills hand off to each other: spark the idea, plan it, build it, break it on purpose, ship it.
Watch an idea ship
One idea. Five skills. Shipped.
2 · /beadsmith
Plan it
Turn an approved plan into trackable work.
3 · /forge
Build it
Take a project from idea to release.
4 · /bug-zero
Break it on purpose
Run a thorough project-wide bug hunt.
5 · /ship-it
Ship it
Prepare a working project for release.
Released
And the crowd goes wild.
That's the whole trip, one skill per step. Take all 52 and run your next idea through it.
The skills, connected
How the skills flow together
39 core skills · 52 connections
Hover to trace. Select to explore.
Plan & shipCode qualityPerformanceDesign & writingAgents & automation
Swipe across the map, or choose a skill below.
Suggested handoffs between skills. Select a node to see what leads into it and where to go next.
Suggested sequence
Choose useful work, organize the implementation, then prepare it for release.
Plan & ship
Idea to release
/forge
Take a project from idea to release.
Try it: Turn a rough habit-tracker idea into a plan, working app, and release.
Explore this skillUseful work before this
Use the gap between the vision and working product to guide the next build.
Where to take it next
Give a new project a working foundation before expanding it.
Carry the project forward through the remaining implementation work.
Review the implementation for defects before calling it ready.
Review the whole user experience as the product takes shape.
Review performance across the implementation.
Every skill, free
Find the skill for your next task.
All 52 skills are free and open source. Open any skill to read its full instructions, or download them all at once.
Showing 52 of 52 skills
- Plan & ship001/052Idea to release/forgeTake a project from idea to release.Handoffs
- Code quality002/052Find bugs across the project/bug-zeroRun a thorough project-wide bug hunt.Handoffs
- Performance003/052Prove the speedup/optimize-proofTurn measured bottlenecks into verified speedups.Handoffs
- Design & writing004/052Refine the finishing touches/polish-passMake an existing interface feel more refined.Handoffs
- Design & writing005/052Make setup easy to follow/readme-realWrite accurate READMEs that sound human.Handoffs
- Agents & automation006/052Repeat useful work/agentic-loopsRun repeatable agent workflows on a cadence or goal.Handoffs
- Code quality007/052Find risky access paths/security-sweepFind and fix high-impact SaaS security risks.Handoffs
- Plan & ship008/052Choose what to build next/idea-forgeTurn project context into a focused improvement backlog.Handoffs
- Plan & ship009/052Turn plans into trackable tasks/beadsmithTurn an approved plan into trackable work.Handoffs
- Plan & ship010/052Build around customer demand/deposit-first-shippingKeep product work tied to real customer demand.Handoffs
- Plan & ship011/052Put ideas through a challenge/dueling-idea-forgeChallenge ideas across multiple AI models.Handoffs
- Plan & ship012/052Start with a working foundation/kickstartTurn a new idea into a working project foundation.Handoffs
- Plan & ship013/052Keep the project moving/project-autopilotKeep a project moving toward its intended outcome.Handoffs
- Plan & ship014/052Compare promises with reality/reality-checkCompare the project vision with what actually works.Handoffs
- Plan & ship015/052Get ready for release/ship-itPrepare a working project for release.Handoffs
- Plan & ship016/052Find a way past the blocker/problem-breakthroughFind a better path through a stubborn blocker.Handoffs
- Code quality017/052Trace dangerous agent actions/agent-exploit-pathsFind dangerous paths from untrusted input to privileged actions.Handoffs
- Code quality018/052Focus a codebase review/audit-lensReview a codebase through a focused quality lens.Handoffs
- Code quality019/052Inspect the code that matters/bug-sweepFind and fix bugs across a chosen code surface.Handoffs
- Code quality020/052Stabilize and prevent repeats/incident-zeroStabilize production incidents and prevent repeats.Handoffs
- Code quality021/052Move systems with confidence/migrate-proofMove systems safely while protecting existing behavior.Handoffs
- Code quality022/052Fix the bug. Keep the scope small./no-slop-bug-hunterFix the actual bug without unrelated changes.Handoffs
- Code quality023/052Make tests worth trusting/test-maxMake a test suite more trustworthy.Handoffs
- Code quality024/052Follow the failure to its source/trace-itMake intermittent failures easier to locate.Handoffs
- Code quality025/052Untangle the code/simplify-proofReduce code complexity without changing behavior.Handoffs
- Performance026/052Find where time goes/hotspot-scoutFind and rank the biggest performance bottlenecks.Handoffs
- Performance027/052Improve performance end to end/perf-maxImprove performance across the whole project.Handoffs
- Design & writing028/052Choose a visual direction/design-intelSet a clear visual direction for a new interface.Handoffs
- Design & writing029/052Improve the whole user journey/ux-maximizerRaise a product's overall UI and UX quality.Handoffs
- Design & writing030/052Find friction and access barriers/ux-scanFind and prioritize usability and accessibility issues.Handoffs
- Design & writing031/052Write like a person/unslopRemove AI-sounding patterns while preserving meaning.Handoffs
- Agents & automation032/052Go another level deeper/deepenApply a skill repeatedly for a deeper result.Handoffs
- Agents & automation033/052Run models on your own machine/local-llm-opsOperate local AI models with fewer surprises.Handoffs
- Agents & automation034/052Keep heavy jobs within capacity/memguardPrevent heavy AI workloads from freezing the machine.Handoffs
- Agents & automation035/052Keep useful project knowledge/memory-forgeBuild a reliable memory system for AI agents.Handoffs
- Agents & automation036/052Recover the context you need/recallBring the right past context into the current session.Handoffs
- Agents & automation037/052Find lessons in earlier work/session-mineFind useful decisions and patterns in past agent sessions.Handoffs
- Agents & automation038/052Make a method reusable/skillifyTurn a solved problem into a reusable skill.Handoffs
- Agents & automation039/052Coordinate agents on one project/swarm-runCoordinate multiple coding agents on one project.Handoffs
- Plan & ship040/052Debate a big decision/advisor-councilHave several advisor personas argue a hard call, then get one synthesized recommendation.Solo skill
- Agents & automation041/052Build an advisor persona/advisor-persona-buildTurn a real person's documented method into a focused decision-making skill.Solo skill
- Plan & ship042/052Onboard a phone-agent client/ai-receptionist-tenant-onboardingAdd a new business to a multi-tenant AI phone agent and prove it is ready for real calls.Solo skill
- Plan & ship043/052Guard scope on a live stream/build-in-public-scope-guardSort viewer suggestions during a build stream and decline the distracting ones politely.Solo skill
- Plan & ship044/052Sell consulting work cleanly/consulting-sales-playbookWrite proposals, follow up and negotiate scope without drifting into unpaid or equity-only work.Solo skill
- Design & writing045/052Prepare a live demo/stream-demo-prepRehearse a demo, clean up your screen and have a fallback ready before going live.Solo skill
- Agents & automation046/052Run the lifecycle on a local model/qwen-forgeDetect a project's stage and run the matching local-model skill, with no API tokens.Local model
- Agents & automation047/052Start a project on a local model/qwen-kickstartScaffold a new project with a small local model: intent, simplest stack, working skeleton.Local model
- Agents & automation048/052Push a project on a local model/qwen-project-autopilotWork through ready tasks one small verified change at a time on a local model.Local model
- Code quality049/052Hunt bugs on a local model/qwen-bug-zeroFind, confirm and fix bugs one at a time with regression tests, on a local model.Local model
- Performance050/052Speed up code on a local model/qwen-perf-maxProfile, optimize the top hotspot and prove behavior is unchanged, on a local model.Local model
- Plan & ship051/052Prepare a release on a local model/qwen-ship-itDo the reversible release prep locally and list every publish step for you to approve.Local model
- Design & writing052/052Fix UI basics on a local model/qwen-ux-maximizerFix concrete, checkable UI issues locally and leave taste decisions to you.Local model
Install
Two commands, then call a skill.
Copy the skills into Claude Code, start a new session, and type the one that fits your task, like /forge.
git clone https://github.com/SubliminalCoding/bilbroswagginz-skills.git
cp -r bilbroswagginz-skills/skills/* ~/.claude/skills/