Claude Skills by colinmollenhour

Conceive and build thoughtful interactive animated visualizations using divergent concept design, deterministic timelines, and accessible web graphics. Use for animated diagrams, architecture or protocol simulations, data-flow explainers, lifecycle visualizations, SVG animations, or Cursor-style technical visuals in React, Vue, Svelte, or vanilla projects.

Scan a codebase with the CCLSP MCP and fix reported diagnostics. Language- and project-agnostic. Use for "fix all the LSP errors", "check the workspace for errors", "fix diagnostics in <path>", or auditing a branch/PR/commit range for new errors. Default scope is the whole codebase, chunked. User can narrow to files, folders, commits, ranges, branches, PRs, or MRs.

Shell out to claude CLI.

Design, implement, or review CLI UX: args, flags, help, output, errors, config, interactivity, and distribution.

Author a ClickUp task from a conversation, plan, review, incident, or spec as CUFM via cup task-sync. Covers the source-fidelity ledger, the house voice for task copy (plain verbs, declarative mood, paragraphs over lists, 5% length budget), rich description structure (banners, toggles, mermaid, tldraw, tables), open-decision blocks (question, AI Recommendation, two alternatives), title conventions, ShipStream Value Stream / Requested By / Bug-type rules, and the post-write audit. Use when cre...

Run OpenAI Codex CLI headlessly for one-shot tasks, code reviews, and native raster image generation or editing. Use when the user mentions Codex, asks to shell out to GPT through Codex, requests `codex exec` or `codex review`, or wants Codex to create a PNG/JPEG, infographic, thumbnail, mockup, or image variant.

Commit, push, and open/update a Github PR or GitLab MR (unless otherwise specified)

Multi-model spec critique — surfaces inconsistencies, gaps, naming issues, and inferior design

Create an implementation guide for a feature in this repo so you can port it to another repo

Augment the current plan with thorough software development planning to create a thorough spec.

Address open comments on a Pull Request or Merge Request related to the current branch

Inspect and resolve git merge conflicts intelligently

Fix failing CI pipeline (GitHub Actions or GitLab CI) for the current branch

Clean up local git branches and stale worktrees (squash-aware; prefers worktrunk when available)

Quick session handoff dump. Generates a portable markdown document from current conversation context for resuming in any AI tool.

Distill the current task, merge request, pull request, ticket, RFC, or recent change set into a single self-contained image-generation PROMPT for a visual infographic that conveys "what changed and why" to non-implementers. Use when the user asks to make/write/draft/design an infographic for an MR, PR, feature, change, ticket, or task — or to attach a visual to a review. The skill emits only the prompt; image generation is left to a downstream tool (Nano Banana, Midjourney, DALL·E, NotebookLM...

Write or update the description of a GitHub PR or GitLab MR - summary, why, changes, exact test plan, AI attribution, and a validated educational brief. Use whenever creating a PR/MR, rewriting its description, or posting a note for new commits on an existing one. Do not use when the user asks for a quick, fast, or simple PR/MR.

Audit progress on the task in scope and keep working until it's actually 100% done — no deferring parts of the spec.

Critique the merits of a merge or pull request, not the typical code review.

Code review for GitHub PRs, GitLab MRs, or any git diff

Report unclosed roborev failing reviews across all tracked projects for a time window (default last 24h), map each branch to its worktree, use git log to judge whether findings are already fixed, and emit a residual-bugs report organized by repo/worktree/branch. Report-only: never write code, commit, close reviews, or switch branches. Use when the user asks for a roborev summary, open/unfixed roborev bugs, unclosed findings across projects, /colin-roborev-summary, or "which roborev bugs are s...

Smart squash-merge a branch onto master, preserving per-author commits with AI-generated messages

Multi-role ultra production-readiness audit — N models × 3 focused roles (hardening / operability / stewardship) on the current repo state

Multi-model, repository-aware bug review with focused discovery, evidence validation, and convergence rounds

Generate a project template for Coolify

Review docs, score issues, preview edits. Use for READMEs, guides, help docs, Markdown/MDX/RST/HTML.

Write or edit user-facing technical docs: READMEs, guides, tutorials, API refs, changelogs, release notes, troubleshooting.

Use Drizzle ORM for TypeScript schemas, queries, migrations, relations, and transactions.

Create a grounded educational brief for delivered PRs, MRs, branches, features, or agent runs. Use when packaging a journey, design decisions, architecture, diagrams, and lessons into a dense reviewer-facing or future-agent-facing brief.

Generate Playwright E2E tests from user stories or manual workflows; explore flow interactively before writing code.

Use gh for GitHub PRs, issues, runs, releases, comments, checks, labels, and raw API tasks.

Manage GitHub Security Advisories: validate reports, draft GHSA/CVE/CVSS fields, patch private forks.

Use glab for GitLab MRs, issues, discussions, pipelines, notes, labels, and raw API tasks.

Shell out to grok CLI (Grok Build TUI) for headless xAI Grok runs.

Generate a self-contained image-generation prompt for an infographic.

Run a multi-agent debate to compare options and converge on a decision.

Run the same task with multiple agents for reviews, critiques, or model comparison.

Autonomous large-task delivery agent. Use for long-running coding work that should go from objective or plan to implemented code, review fixes, PR/MR, and green CI with minimal human-in-the-loop gates.

Generate or edit images with Nano Banana/Gemini CLI: thumbnails, icons, diagrams, graphics, illustrations, photos.

Author or edit Nuxt Content Markdown using MDC syntax and Vue components.

Build or debug UI with Nuxt UI components, forms, tables, modals, and current docs.

Act as a brand-new user of a product and audit its first-run onboarding end to end — sign up with a real email, drive the web app / desktop app / CLI, log every dead end, unexplained error, and undocumented prerequisite, then write a severity-ranked friction report with concrete fixes. Use when asked to "test onboarding", "sign up as a new user", "dogfood the first-run experience", "onboarding QA/audit", or to reproduce a new-user journey and report the friction.

Audit app security beyond checklists: abuse prevention, API protection, business logic, rate limits, validation.

Create or improve Agent Skills, SKILL.md frontmatter, structure, instructions, and resources.

Create or configure VoltAgent agents with tools, memory, hooks, sub-agents, and workflows.

Sketch explanations on a live-reloading HTML whiteboard the user watches in their browser while you edit. Use when the user asks for a whiteboard, asks you to "draw", "sketch", "diagram", or "walk me through" a design, change, architecture, flow, schema, or plan visually, or when a running explanation of 3+ interacting parts would read better as a page than as terminal text. Not for polished published pages (use Artifacts) or static image exports (use tldraw-skill / pretty-mermaid).

Shell out to claude CLI.

Shell out to codex CLI.

Create a grounded educational brief for delivered PRs, MRs, branches, features, or agent runs. Use when packaging a journey, design decisions, architecture, diagrams, and lessons into a dense reviewer-facing or future-agent-facing brief.

Use gh for GitHub PRs, issues, runs, releases, comments, checks, labels, and raw API tasks.