skills

Audits and fully configures how a website looks when shared (Open Graph, Twitter/X cards, WhatsApp/Discord/Slack previews) and in the browser (title, favicons, theme-color, web manifest), generating missing assets and verifying the result locally. Use when setting up link previews, og:image, favicons, social sharing metadata, or when a shared link "looks wrong" on a platform.

Iterative React-health loop: inventory every React app in the repo, baseline with the latest react-doctor (pinned for the rest of the run), fix one root cause per iteration, and re-verify until every app scores 100/100 or every remaining finding has a user-approved written justification. Never silences findings with ignores, exclusions, or deleted behavior, and stops itself after two iterations without measurable progress. Use when the user mentions react-doctor, React health score, React anti-patterns cleanup, or driving a React codebase to a clean scan.

Iterative dependency-CVE remediation loop: scan with ecosystem-native advisory tooling, assess whether each high/critical finding is actually reachable (with call-path evidence, not vibes), fix the highest-risk reachable one with the smallest credible change, re-verify, and repeat. Terminates when no exploitable high/critical CVE remains or every remaining finding has an evidence-backed reachability assessment and an approved risk decision. Use when the user mentions CVEs, vulnerability scanning, dependency security, npm/pip/cargo audit, security patching, or wants a recurring dependency-security loop.

Orchestrates a comprehensive frontend UI review by running parallel analysis across typography, layout, accessibility, responsiveness, copy, visual polish, and design critique. Uses browser inspection for visual context. Produces a prioritized plan of small, concrete fixes -- never rewrites. Use when the user mentions UI review, frontend review, page review, design review, visual review, audit my page, or review my UI.

Review open PRs authored by the user for a given repo, showing title, status, review state, open comments, CI checks, and recommended next steps. Use when the user mentions PR review, my PRs, open PRs, PR status, PR dashboard, or wants an overview of their pull requests.

Surfaces the single most important blindspot in a plan, decision, or situation by checking whether the frame is right, generating candidates across 8 lenses (hidden constraints, false assumptions, second-order effects, stakeholders, blast radius, base rates, time horizon, counter-positions), then either committing to one or returning a calibrated "no blindspot" when none is load-bearing -- never a list of risks. Use when the user asks "what am I missing", "what's the most important thing I'm missing about this situation or decision", "what's my blindspot", "what's the biggest risk I'm not seeing", or wants a single sharp second opinion before committing.

Read-only audit of a web page's layout grid. Measures column adherence, baseline rhythm, and optical ink alignment in pixels (Müller-Brockmann), then critiques grid appropriateness, type scale, flush-left, white space, measure, and color-as-identifier (Vignelli). Mechanical px violations first, principled critique second, each with a minimal fix -- never rewrites. Works rigorously on pages with a declared CSS grid; falls back to inferred-grid plus principled checks on arbitrary pages. Use when the user mentions grid review, grid audit, layout grid, column alignment, baseline grid, is my grid aligned, Swiss grid, Müller-Brockmann, or Vignelli.

Bootstrap a complete agent guide system for a repo — generates AGENTS.md (always-on), .agents/{domain}.md (loaded on demand), CLAUDE.md (redirect), and .agents/design-decisions.md using progressive disclosure and canonical references instead of stale embedded code. Use when the user mentions setting up AGENTS.md, bootstrapping an agent guide, configuring CLAUDE.md for a repo, "set up agent context", or onboarding a new repo for AI coding agents.

Land finished worktree work onto main for solo projects: auto-commit pending changes, sync with origin/main resolving conflicts inside the worktree, fast-forward main, push, and remove the worktree so the conversation can be archived. Use when the user says save and archive, land this worktree, ship this to main, push this conversation's work to main, guardar y archivar, or says they're done with a worktree conversation and want the work saved on main before closing it.

Teach the user a new skill or concept across multiple sessions, using the current directory as a stateful teaching workspace (mission, curated resources, learning records, glossary, and beautiful HTML lessons). Use when the user says "teach me X", wants to learn a topic over time, or asks for a personalized lesson plan.

Generate a Slack-ready governance proposal message from a single proposal URL (forum post, Tally vote, Snapshot, etc.). Fetches the page, extracts proposer, summary, context, and risk, then drafts the message using the fixed governance template. Use when the user mentions governance message, gov proposal, gov message, gobernanza, Slack message for a proposal, ZKsync proposal summary, Tally vote summary, or pastes a proposal link and asks for a write-up.

Read or update the PR description for the current branch, strictly following the repo's PR template. Enforces concise bullet points and template compliance. Use when the user mentions PR description, PR body, updating a pull request, or filling in the PR template.

Interviews the user relentlessly about a plan or design until reaching shared understanding, walking the decision tree one branch at a time. Use when the user wants to stress-test a plan, get grilled on their design, or mentions "grill me".

Performance regression detection for web pages -- captures baselines for page load, Core Web Vitals, and resource sizes, then compares subsequent runs and flags regressions by configurable thresholds. Use when the user mentions benchmark, perf check, performance regression, Core Web Vitals, page speed, or wants to track frontend performance before and after a change.

Generate a short, copy-paste friendly Slack message listing the user's open PRs in the current repo, split into "Ready for review" (non-draft) and "Drafts" sections, each as bullets with custom emoji and markdown-style links. Use when the user mentions PRs ready to review, Slack PR list, share my PRs on Slack, or wants a short message to ping reviewers.

Set up and manage git worktrees so each agent conversation works on its own branch in its own physical directory. Prevents parallel agent chats (Cursor, Claude Code, etc.) from clobbering each other's git state. Use when the user says they're starting a new issue, starting an implementation, picking up a ticket, working on a new branch, or mentions worktrees, parallel branches, branch isolation, working on multiple branches at once, or when agents have gotten confused about which branch they're on. Also use when the user says a PR is done/merged and wants to clean up its worktree.

Sets a per-workspace completion condition and keeps Cursor auto-iterating across turns until a small-model evaluator confirms the condition holds. Modeled on Anthropic's Claude Code /goal command but adapted to Cursor's stop hook. Use when the user types /goal-cursor, /goal-cursor status, /goal-cursor clear, /goal-cursor stop, or asks to set / show / clear a long-running goal in Cursor.

Critique and rewrite a prompt using prompt engineering best practices: clarity, examples, XML structure, role, explicit output format, and positive-over-negative instructions. Asks 1-3 targeted questions to fill missing context, then returns a short critique plus a drop-in rewritten prompt. Use when the user wants to improve, refine, audit, fix, or rewrite a prompt.

Systematic root-cause debugging that enforces an "Iron Law" (no fix without a confirmed root cause) and walks investigation through five phases -- collect, pattern-match, hypothesize, fix, verify. Use when the user says debug this, fix this bug, why is this broken, root cause, or investigate.

Build Meta Ads Manager bulk-import files (.xlsx + Unicode .txt) from a structured YAML brief, using the bundled v2.3 Ads Manager Template. Expands a creative matrix (titles x bodies x creatives) into one row per ad and validates every field against Meta's enums and 2026 character limits. Use when the user mentions Meta ads bulk, Facebook ads bulk, Ads Manager template, .xltx, bulk import, mass ad creation, AdsManagerTemplate, or wants to launch many ads at once.

YC-style office hours brainstorm that interrogates the premise of a product idea (or builder side-project) through six forcing questions, generates 2-3 alternative approaches, and ends with a written assignment -- never writes code. Use when the user mentions office hours, brainstorm this, is this worth building, validate this idea, or wants a sharp thinking partner before they start coding.

Review a teammate's PR and produce a small list of high-signal, senior-engineer-level review comments ready to paste into GitHub. Aggressively drops nitpicks, style preferences, and "would be nice" suggestions. Output is text-only -- never auto-posts. Use when the user mentions review-pr, reviewing a teammate's PR, leaving PR comments, or wants senior-level feedback to post.

Drives a local web app through Chrome MCP across the happy path plus 2-3 edge cases, capturing screenshots, console errors, and network failures to produce QA evidence. Use when manually QA'ing a web feature, testing a flow in the browser, exercising a change before claiming it works, or producing evidence for verification-before-completion.

Forces running a fresh proof command and reading its output before any "done" / "fixed" / "works" / "passing" / "ready to commit" / "ready to merge" claim. Use when about to assert that a change succeeded, before committing, before opening a PR, or before marking a task as finished.

Scans the current GitHub repo for open, unassigned issues with no linked PR, scores them against an AI-readiness rubric, and returns the top 3 picks with rationale and a ready-to-paste agent kickoff prompt for each. Use when the user mentions AI triage, AI-ready issues, picking issues for an agent, what to work on next, or wants to find issues to dispatch to an AI.

Author or audit agent skills: gather requirements, write SKILL.md with frontmatter, structure supporting files, and verify quality. Use when creating a new skill, reviewing an existing skill, or asking about SKILL.md format and best practices.

Pull the latest changes from main and merge them into the current working branch. Use when the user mentions updating their branch, syncing with main, pulling latest, refreshing from main, or merging main into their branch.

Heavyweight PR verification that produces a structured report: code review (delegates to miguel-review), test verification with gap analysis and execution, and upstream assumption validation. Use when the user mentions verify PR, verify my PR, PR verification, test plan for PR, or validate PR changes.

Create a GitHub issue in the current repo using one of the predefined issue templates (bug, feature, tech debt) or as a blank issue. Fetches available labels dynamically, maps user intent to the right template, and enforces concise problem descriptions. Use when the user mentions creating an issue, filing a bug, opening a ticket, or reporting a problem.

Fetch the latest PR review comments for the current branch, analyze and summarize feedback, and help the developer plan fixes. Use when the user mentions PR comments, review feedback, reviewer requests, or wants to understand and address PR review comments.

Opinionated code review that prioritizes minimal diffs, deletions over additions, and zero tolerance for dead code or premature abstractions. Reviews the current branch diff against a base branch. Use when the user mentions miguel review, PR review, diff review, code review, or audit.

Write or rewrite content in a direct, personal, sensory style that feels human. Applies voice principles, bans AI-giveaway phrases, and enforces visual formatting rules. Use when writing blog posts, social media, website copy, emails, PR descriptions, commit messages, or any written content. Use when the user asks to write, rewrite, edit copy, review tone, or match their voice.