klimkit

Convert an implementation, debugging, release, UI, backend, security, or research request into a concrete pass/fail acceptance checklist before code changes start. Use when Codex needs blocking success criteria, scope boundaries, verification targets, or a task note for Klimkit-managed work.

Diagnose bugs, failures, flaky behavior, broken tests, CI failures, or unexpected production signals with a reproduce-first workflow and Klimkit proof. Use when the user asks to debug, investigate, root-cause, fix a failure, or explain why behavior changed.

Interview the user relentlessly about a plan, design, architecture, product decision, workflow, or release choice until shared understanding is reached, using repo and web research to prioritize the most important unresolved questions while recording the considered question shortlist, each asked question, and each approved decision in a Klimkit task note. Use when the user asks to be grilled, stress-test a plan, resolve design branches, clarify tradeoffs, or turn vague intent into explicit decisions.

Run Klimkit's full implementation workflow for code, docs, release, configuration, UI, backend, security-sensitive, or proof-producing changes. Use when the user asks Codex to implement, fix, refactor, release, migrate, wire up, or complete non-trivial repository work with checklist, verification, reflection, and final review.

Write a concise cross-task reflection note after verification and before final review for non-trivial Klimkit work. Use when Codex should connect the current task to prior work history, patterns, risks, or future process improvements.

Check, explain, or set up Klimkit static proof-report serving for `docs/work` HTML reports and Tailscale URLs. Use as the fallback when native hosted publishing (Claude Code Artifacts, Codex Sites) is unavailable and a task needs report-server readiness, `/reports/` links, Tailscale Serve verification, or troubleshooting report visibility.

Set up Klimkit's docs-first repo context for agent work. Use when adding the `docs/work/` journal layout, repo skill pointers, config state, or onboarding instructions to a project that should use Klimkit workflows.

Implement behavior changes with narrow red-green-refactor loops and Klimkit proof. Use when adding features, fixing bugs with clear expected behavior, changing shared contracts, or when the user asks for TDD or test-first work.

Create human-readable Klimkit walkthrough proof reports with steps, screenshots, links, redaction checks, and report-server/Tailscale handoff. Use when a user asks for a walkthrough, proof report, demo evidence, QA narrative, or review packet.

Audit and clean Codex and Claude Code customization across local computers, SSH hosts, containers, and VMs with a two-phase approval gate. Use when the user asks to inventory, compare, reset, consolidate, quarantine, remove, or troubleshoot agents, skills, instructions, hooks, MCP servers, plugins, worktrees, caches, or harness configuration on one or more machines.

Drive OpenAI Codex through the `codex` CLI and Codex desktop app from any agent or script. Use when Codex needs to spawn sessions, resume or connect to existing sessions, send live desktop messages, read transcripts, orchestrate many Codex agents, use the Codex app-server WebSocket protocol, or find active Codex sessions.

Create isolated Git worktrees for Klimkit agent work, including simple branch worktrees and feature branches from a base branch that is first updated by merging another branch. Use when the user asks to create a worktree, make a new branch checkout, start isolated feature work, or prepare a dev-synced worktree.

Review auth, authorization, secrets, data exposure, input validation, sandboxing, infrastructure, compliance-sensitive flows, and unsafe defaults before completion. Use when a change touches security boundaries, credentials, deployment, containers, permissions, user data, or network exposure.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks.

Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.

UI/UX design intelligence for web and mobile. Includes 50+ styles, 161 color palettes, 57 font pairings, 161 product types, 99 UX guidelines, and 25 chart types across 10 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, and HTML/CSS). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, and check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, and mobile app. Elements: button, modal, navbar, sidebar, card, table, form, and chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, and flat design. Topics: color systems, accessibility, animation, layout, typography, font pairing, spacing, interaction states, shadow, and gradient. Integrations: shadcn/ui MCP for component search and examples.

Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my site against best practices".

Perform a fresh-context final acceptance review of the exact draft response against the original request, checklist, changed files, verification evidence, reflection, and proof artifacts. Use as the last gate before telling the human work is complete; especially use from fresh subagents when `klimkit-implement` requests final review.

Trace repository behavior across files before implementation without making edits. Use when Codex needs to understand unfamiliar code, execution paths, architecture, dependencies, tests, ownership boundaries, or contradictory patterns before deciding what to change.

Browser automation through the agent-browser CLI for websites, local web apps, forms, screenshots, scraping, QA, and exploratory testing. Use when Codex needs real browser interaction, visual evidence, accessibility snapshots, video recording, or browser-based checks.

Distinctive, production-grade frontend design for web components, pages, applications, dashboards, and interactive tools. Use when Codex needs to build or revise a user-facing UI with strong visual direction, polished interaction states, and non-generic aesthetics.

Structured plan interrogation that asks one sharp question at a time until the design is clear. Use when the user says "grill me", asks to stress-test a plan, or needs dependencies, assumptions, and tradeoffs surfaced before implementation.

Safe edits to Klimkit's repo-managed Codex harness pack, including shared AGENTS guidance, model config, subagents, hooks, and bundled skills. Use when the user asks to change `packs/codex/` behavior or `kk apply` projections, not for normal plugin-only workflow use.

Klimkit's evidence-first Codex workflow for implementation, debugging, review, and release tasks. Use when a user asks for Klimkit, wants checklist/proof-driven work, needs `.klimkit` task artifacts, or expects verification, reflection, and final review before handoff.

Browser automation CLI for AI agents. Use for websites, local web apps, forms, clicks, screenshots, scraping, QA, and exploratory testing. Also applies to Electron app automation, Slack checks, cloud browsers, and local browser tasks.

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.

Interview the user relentlessly about a plan or design until shared understanding is reached. Use when they want a plan stress-tested or mention "grill me".

Tune __HUMAN_NAME__'s shared Codex home-level harness pack safely through the Klimkit repo, not by editing generated files in ~/.codex directly.