Engineering & prompt-craft skills
Developer-flavored skills that work in Cowork & Chat — code review, clean-code standards, security review, prompt engineering, and framework best-practices. Advisory and file-based, no live server needed. Click any skill for its full guide with 20 Melinda Maria examples.
code-reviewerCode ReviewerCowork & Chat
A thorough second set of eyes on code — it reviews TypeScript, JavaScript, Python, Swift, Kotlin, and Go for bugs, quality issues, and security problems, and produces a clear checklist. Perfect for safely vetting AI-written code before it goes live.
e.g. "Review this new tool Claude wrote for our canto-mcp connector before I deploy — flag bugs, security issues, and anything that could leak data, as a checklist."
clean-codeClean CodeCowork & Chat
Pragmatic coding standards that keep AI-written code simple, readable, and not over-engineered — clear names, small functions, no dead features. It makes our connector code easier for a non-coder to maintain later.
e.g. "Clean up this connector tool Claude generated — simpler names, smaller functions, drop anything we don’t actually use — so I can still read it in six months."
senior-securitySenior SecurityCowork & Chat
A security-engineering advisor — application security, threat modeling, and audits. It helps us reason about how our connectors and data could be attacked and what to lock down, in plain English.
e.g. "Threat-model our custom MCP connectors — they sign in with our MM Google work account and some hit write-capable APIs — and tell me, plainly, where the biggest risks are."
senior-prompt-engineerSenior Prompt EngineerCowork & Chat
Expert help writing better prompts and instructions for Claude — the patterns, structure, and few-shot examples that make answers more reliable. Directly useful for sharpening the instructions in our skills, connectors, and Projects.
e.g. "Help me write the instructions for a new MM skill so it always replies in our brand voice and returns a blog post + a Klaviyo excerpt + 3 captions in a consistent format."
vercel-react-best-practicesVercel React Best PracticesCowork & Chat
Vercel’s deeper companion guide — 70 React/Next.js performance rules across 8 categories, ranked by impact. Use it for a more exhaustive performance pass on a React codebase than the core React guide.
e.g. "Run a full Vercel-rules performance review on this Next.js page and group the findings from critical to low."
github-actions-docsGitHub Actions DocsCowork & Chat
Docs-grounded help for GitHub Actions — the automation that runs when you push code (e.g. auto-test or auto-deploy). It answers from the official GitHub docs instead of stale memory, so the YAML it gives you is current. Relevant if we put a connector repo on GitHub with auto-deploy.
e.g. "Write a GitHub Actions workflow that auto-deploys a connector to Cloudflare with wrangler whenever I push to main, and link the official docs for each part."
grill-meGrill MeCowork & Chat
A practice / quiz coach: it tests you on a topic with questions so the material actually sticks — a personal learning tool rather than a connector or build skill. (Details kept high-level; the skill’s file wasn’t readable in this build, so this describes its purpose only.)
e.g. "Grill me on how our custom connectors work — ShipBob vs ShipHero, read-only vs write-gated — so I can confidently explain them to the team."
