# Digital Alchemy: Claude Code Skill Packs > Every skill listed here is a real skill used in production builds at https://digitalalchemy.dev. Packs are sold individually; each pack has at least one free skill. Human-readable catalog: https://digitalalchemy.dev/skills ## 01 Setup Set up how you work before you build anything. These skills decide how much your AI does on its own, and which AI model handles which job. Get this pack: https://digitalalchemy.dev/api/stripe/checkout/setup ($12) ### /skill-map (free) See your whole skill library, graded. Turns every skill you own into a set of web pages you can actually read. Each card shows what the skill does, the exact words that wake it up, which of your AI tools have it, and how often you really use it. It grades each one and flags any two skills fighting over the same trigger words. You also get a spreadsheet of all of it. When to use: - You have collected more skills than you can remember - A skill never fires and you cannot work out why - Two skills keep triggering on each other and you need to see which words overlap - You want to know which skills earn their keep and which are dead weight Detail page: https://digitalalchemy.dev/skills/skill-map ### /windows-ps51-guardrails (free) Stop Windows from quietly breaking your builds. Teaches your AI the Windows rules that burn hours. It stops writing commands your version of PowerShell cannot run. It stops saving settings files in the format that corrupts them. Same request, commands that actually work. When to use: - You build on Windows and commands keep failing for no clear reason - A settings file saved fine but the app cannot read it - Your AI keeps writing commands your shell rejects Detail page: https://digitalalchemy.dev/skills/windows-ps51-guardrails ### /output-style-lock (free) Make the AI talk the way you actually read, permanently. Writes one saved line that changes how every future answer comes back to you. Short sentences, plain words, and no jargon you have to go look up, plus a second line so it still sounds like a person and not a manual. If the ready made plain English standard is not your taste, it asks you nine questions and builds your own style instead. You also get five quick questions to grade any answer against your rule, so you can tell when it slipped instead of guessing, and a way to keep the same rule identical across Claude, ChatGPT, and Gemini instead of three copies quietly drifting apart. There is an off switch for the times you want the rule ignored. When to use: - Answers keep going over your head and you read them twice - You retype 'explain it simply' in every single new chat - You want one style rule that works in Claude, ChatGPT, and Gemini Detail page: https://digitalalchemy.dev/skills/output-style-lock ### /pm-director-mode Decide how much your AI builds before it checks in with you. Sorts every task into three levels. Small fixes ship on their own. Medium work gets a heads up. Big changes wait for your yes. You stop being asked about tiny things, and your AI stops rebuilding half the app when you asked for a button. When to use: - Your AI asks permission for changes that do not need it - You asked for one button and got a rewritten app - You want the same working rhythm in every session Detail page: https://digitalalchemy.dev/skills/pm-director-mode ### /model-router Send each job to the cheapest AI that can still do it well. Sorts work by how many judgment calls it needs, not by how hard it sounds. Real thinking goes to the expensive model. Repetitive output goes to the cheap one. Your bill drops and the quality holds. When to use: - Your AI spending is climbing and you cannot say why - You run background helpers and must pick a model for them - You have a big batch of small tasks to run Detail page: https://digitalalchemy.dev/skills/model-router ### /branch-triage Clean up old work-in-progress copies without losing anything. Checks every saved copy of your project, proves which ones are already folded into the main version, and deletes only those. You end with a clean project and a short list of the decisions that were actually yours to make. When to use: - You are scared to delete old copies in case something is in them - You cannot remember whether everything is saved and backed up - You worked on one project across two machines Detail page: https://digitalalchemy.dev/skills/branch-triage ## 02 Plan-to-Ship Flow The loop that takes an idea all the way to a live product. Build fast, put it online safely, watch what people actually do, and keep what you learned. Get this pack: https://digitalalchemy.dev/api/stripe/checkout/plan-to-ship ($20) ### /railway-deploy-checklist (free) Put your app online without the 2 a.m. surprise. Walks your launch through the checks that catch problems while you are awake. It confirms every setting loads before the app starts. It adds a page that reports whether the app is healthy. It writes your undo steps before you need them. When to use: - You are putting an app or tool online for real users - It works on your machine and dies once it is live - You want your undo plan written down before you launch Detail page: https://digitalalchemy.dev/skills/railway-deploy-checklist ### /done End every session so the next one starts where you left off. Closes a work session by collecting what changed, writing it up, saving it to your notes, and updating what your AI remembers. Tomorrow you skip the twenty minutes of re-explaining the project. When to use: - You are wrapping up and want the context kept - You keep re-explaining the same project to your AI - You want a searchable record of what shipped and why Detail page: https://digitalalchemy.dev/skills/slash-done ### /generator-app-clone Launch a small AI tool from a template that is already safe. Copies a proven setup: one AI call that runs on your server where nobody can steal your key, strict checks on anything a user types, limits so nobody can hammer it, and a deploy. You get a finished product in hours because the safety work is done. When to use: - You want to ship one focused AI tool fast - You are launching several products that share one shape - You want the safety baseline without building it again Detail page: https://digitalalchemy.dev/skills/generator-app-clone ### /funnel-telemetry-ship See where people quit before you build the next feature. Adds five checkpoints to your app: signed up, got started, read the guide, took the first action, finished it. Then it shows you how many people make it through each one. The step where people drop off picks your next job, instead of a guess. When to use: - You shipped something and keep guessing what to build next - You have users but no idea where they give up - You are about to start something new instead of fixing this Detail page: https://digitalalchemy.dev/skills/funnel-telemetry-ship ### /daily-debrief Roll every session you closed today into one summary. Reads the day's session notes and pulls out the wins, the blockers, the lessons, and the patterns that keep repeating. Then it files each new idea where it belongs. Nothing you worked out today gets lost. When to use: - You work several sessions a day across different projects - Good ideas keep dying in old chat scrollback - You want a real record of what each day produced Detail page: https://digitalalchemy.dev/skills/daily-debrief ### /retro-debrief Write the session notes you never got around to. Goes back through your old meeting notes and AI sessions, finds the work you never wrote up, drops the junk, and files the rest into your notes. The months you cannot remember become months you can search. When to use: - You started keeping notes late and the backlog bothers you - Your notes are spread across five tools that disagree - You want your AI's memory to match what really happened Detail page: https://digitalalchemy.dev/skills/retro-debrief ## 03 Security Gates The checks you never skip. Every one of these exists because an AI-built app broke without it, and the fix cost more than the check would have. Get this pack: https://digitalalchemy.dev/api/stripe/checkout/security-gates ($20) ### /vibeaudit-workflow (free) Scan your app for the mistakes AI-written code keeps making. Runs Vibe Audit, the free scanner Desi built, with the right settings. It explains every problem it finds in plain words and writes the prompt that fixes it. It looks for the specific ways AI-built apps fail, which normal code checkers do not test for. When to use: - Before you launch anything you built with AI - You took over a project and need to know the risk - You want a safety check that blocks bad code automatically Detail page: https://digitalalchemy.dev/skills/vibeaudit-workflow ### /supply-chain-gate Check any add-on before you let it onto your machine. Treats installing someone else's code as a security event. Before it lands, this checks the package against known attacks, reads any setup script that runs automatically, and watches for behavior that looks wrong. Brand new malicious packages get caught this way, because nobody has reported them yet. When to use: - Before you install anything you did not write - A tutorial tells you to install a package you have never heard of - You are choosing between two add-ons and want the safe one Detail page: https://digitalalchemy.dev/skills/supply-chain-gate ### /clone-and-run-gate Decide whether a stranger's project is safe to download and run. Investigates the project and the account behind it before you run their code. It checks how old the account is, whether the stars look bought, whether any add-on comes from somewhere unexpected, and what runs automatically on install. Then it installs in a sandbox first. When to use: - A project is trending and you want to try it today - A client hands you code and you do not know where it came from - The setup steps include something you do not understand Detail page: https://digitalalchemy.dev/skills/clone-and-run-gate ### /stripe-webhook-hardening Lock down the code that hears back from your payment provider. When someone pays, Stripe calls a page on your site to tell you. Anyone can pretend to be Stripe and call that page. This applies the full lockdown: prove the message really came from Stripe, ignore repeats, and work out the amount owed on your own server instead of believing what the message claims. When to use: - You are adding payments to anything - Your AI wrote the payment code and nobody has checked it - You want the money math confirmed on your side, not the customer's Detail page: https://digitalalchemy.dev/skills/stripe-webhook-hardening ### /secret-leak-triage The calm, correct response to a leaked password or key. Checks whether the leak is real before you panic. It searches your whole project history, not just today's version, because a deleted key is still in there. Then it walks you through replacing the key at the source and installs a guard so it cannot happen again. When to use: - Something flagged a possible password or key in your project - You pasted a key somewhere you should not have - You want an automatic check that catches this before you save Detail page: https://digitalalchemy.dev/skills/secret-leak-triage ### /drizzle-rls-supabase Build your database locked by default, not patched later. Sets up your database so nothing can read or write anything unless you explicitly allow it. Only your own server code gets through. Most leaks happen because a database was open by default and nobody noticed until strangers were reading customer records. When to use: - You are creating a database or adding a table to one - Your app talks to the database straight from the browser - You need the permission rules written correctly the first time Detail page: https://digitalalchemy.dev/skills/drizzle-rls-supabase ## 04 Content Engine Turn what you build into content that keeps working. Community posts, social copy, long articles, and a check that stops you giving away paid material for free. Get this pack: https://digitalalchemy.dev/api/stripe/checkout/content-engine ($12) ### /social-media-writer (free) One idea in, real posts for seven platforms out. Takes a topic, a screenshot, or a rough thought and writes the actual posts for Instagram, TikTok, LinkedIn, YouTube, Facebook, X, and Threads. Each one is written the way that platform expects, instead of one caption pasted seven times. When to use: - You shipped something and want it announced everywhere - You have ideas but posting everywhere takes your whole afternoon - You want posts that fit each platform, not one recycled caption Detail page: https://digitalalchemy.dev/skills/social-media-writer ### /opus-director (free) Give your content tools real direction instead of a vague ask. Writes precise instructions for the AI tools that cut your clips and schedule your posts. Instead of a vague request that returns generic output, you hand the tool an angle, an audience, and the exact moment that matters. When to use: - Your clip and scheduling tools keep giving you generic results - You are writing instructions for another AI writing tool - You want campaign copy directed, not improvised Detail page: https://digitalalchemy.dev/skills/opus-director ### /skool-post-writer Turn a screenshot or half-formed idea into a finished community post. Takes whatever you drop in and turns it into a post with an opening that stops the scroll, a middle that teaches something, and a matching cover image prompt. Built for people who have to post to a community every day. When to use: - You run a community and need posts that stay consistent - You have the insight but not the hour to write it up - You want cover art planned with every post Detail page: https://digitalalchemy.dev/skills/skool-post-writer ### /substack-post-writer Long articles that keep their formatting when you paste them in. Writes the whole article in a format that survives the paste into Substack. Headings, pull quotes, buttons, image prompts placed where the images go, and the search and preview text, all in one piece. When to use: - You write a newsletter and drafting eats your week - Your formatting falls apart every time you paste it in - You want image prompts sitting where the images belong Detail page: https://digitalalchemy.dev/skills/substack-post-writer ### /ip-guard Catch yourself giving away the paid material before you post it. Reads anything you are about to publish and sorts it three ways: safe to show, tease only, or keep behind the paywall. Your method and your paid material get flagged before you post, not after someone screenshots it. When to use: - You sell a program and also post free content about it - A draft explains your whole method step by step - You want one last check before you publish Detail page: https://digitalalchemy.dev/skills/ip-guard ## 05 Client Work The whole consulting pipeline. First call, deep audit, branded report, fast prototype, funnel build, and a daily routine that works your real leads. Get this pack: https://digitalalchemy.dev/api/stripe/checkout/client-work ($20) ### /parallel-audit-lenses (free) Three specialist reviewers on one target, all at once. Sends three separate reviewers at the same target at the same time: one for safety, one for usability, one for the market. None of them sees the others' notes. When two of them flag the same thing on their own, that is your most trustworthy finding. When to use: - One big review keeps missing what a focused pass catches - You want findings confirmed twice, not just listed - A review has to be finished tonight Detail page: https://digitalalchemy.dev/skills/parallel-audit-lenses ### /client-discovery-intake Run a first client call that finds the real problem. Structures the first call so the bottlenecks, the repetitive manual work, and what the client actually wants all surface before anyone builds anything. You leave with a needs list, a scope in phases, and a price to anchor against. When to use: - You have a first call with someone who might hire you - Clients keep asking for features instead of results - You want scope and price settled before you start Detail page: https://digitalalchemy.dev/skills/client-discovery-intake ### /client-audit-deliverable A full site review that ends in a report clients keep. Runs the whole review: scans the site, checks it three ways at once for safety, usability, and market fit, grabs screenshots as proof, and builds a branded report. The report is the product, and it sells the work that comes next. When to use: - A prospect needs to see the problems before they buy the fix - You want reviews that take an evening, not a week - You need something that looks like it cost real money Detail page: https://digitalalchemy.dev/skills/client-audit-deliverable ### /client-prototype-sprint Go from review to a working demo in 48 hours. Picks the client's single most painful task and builds just that one screen, using fake data so it moves fast. Then it puts the demo online and hands you the walkthrough. The pitch writes itself: 48 hours against six weeks. When to use: - A first call went well and you want to close it - The client cannot picture the product from a document - You want to sell the build by showing the build Detail page: https://digitalalchemy.dev/skills/client-prototype-sprint ### /client-funnel-build The customer path you end up rebuilding for every client. Builds the standard path: a free thing worth having, a short form, delivery split by who they are, a run of follow-up emails, and one clear next step. Includes the trick that turns one recording session into weeks of clips. When to use: - A client has an audience but no way to capture or follow up - You keep designing the same path from scratch per client - You need the checklist for a path that is underperforming Detail page: https://digitalalchemy.dev/skills/client-funnel-build ### /money-sprint A daily sales routine driven by your real list of leads. Reads your actual lead list, finds the follow-ups you owe, the quotes nobody answered, and the people who went quiet. Then it fills your day with names, where to reach them, and the exact opening line. It ends by scoring what moved you closer to money. When to use: - You do client work but reaching out keeps sliding - Follow-ups die quietly in your pipeline - You want the day's outreach picked by data, not mood Detail page: https://digitalalchemy.dev/skills/money-sprint