claude-starter

Plugin claude-starter from marketplace claude-starter in ryanportfolio/Harness-Firmware. Commands read from the repo's manifest — today.

Skills (56)

• addskill

  Create, import, update, or install repo or personal skills: runtime ownership, resources, discovery checks.

• adopt-repo

  Mirror external repo privately under user's account + overlay firmware: clone, strip template-only files, privacy-sweep, init-project. Use: 'pull this repo into our firmware'.

• advocate

  Only on explicit $advocate: challenge a just-made change before it lands. Never from natural-language requests.

• apply-firmware

  Bring current folder (plain folder, existing repo, old or partial firmware copy) up to full Harness Firmware in place. Mirror external repo → adopt-repo.

• arena

  Parallel attempts at one task, blind judging, best ideas grafted onto strongest. Use: 'try a few approaches', 'build me options', bakeoffs, competing versions, stalled long-horizon or wow-loop step.

• astra-review

  Codex CLI review on gpt-6-astra, medium; codex-review lifecycle. Use: 'have Astra review this'.

• babysit-ci

  Watch PR checks, fix failures: 'watch CI', 'fix CI', 'get checks green', PR with failing or pending checks. Not a bare merge request.

• brainstorming

  Product, interface, workflow, or architecture goals with unresolved material tradeoffs. Skip routine work + implementing an approved design.

• bro

  Plain-language restatement: 'plain english', 'dumb it down', 'I don't understand', 'what does that mean'.

• cache-loop

  Caching or stale content.

• caveman

  Every session reply: terse Caveman prose + built-in Unslop. User-facing deliverables → Writing.

• claude-review

  Claude or Fable asked to review Codex-written code, or cross-vendor review via Claude CLI.

• climb-loop

  Hill-climb a target round after round: Opus, Sol and Astra rank our UI blind against named reference apps (none → redesign-concepts, build the pick, climb past our own versions) until all rank ours first, or a scripted eval climbs to perfect. Use: 'make our UI beat <app>', 'climb our UI without a re

• codex-review

  Second opinion: Codex CLI (gpt-6.1-sol, high) reviews PR, branch, commit, or uncommitted diff; each finding verified. Use: 'have Codex/Sol review this'.

• conversion-ux

  Audit, design conversion flows stage by stage: each element asks easier, honest question. Paywalls, pricing, trials, sign-up, onboarding, checkout, forms, filters, booking. Use: 'why doesn't this convert', 'audit this flow'. Visual concepts → redesign-concepts; copy polish → writing.

• dare

  First principles, questioning the problem: 4 fresh stages decompose, audit, recombine, test; immutable goals + constraints kept.

• deep-plan

  $deep-plan: interview a loose idea in short decision rounds; stop for explicit go before anything is built.

• design-prototypes

  $design-prototypes: brief → distinct image concepts (sections, features, branding) → compare → refine pick. Before build; built UI → redesign-concepts.

• enhance-prompt

  User wants a rewritten, copy-ready prompt for another agent or session. Produce prompt; don't run its task.

• external-review

  Explicit $external-review, or a launcher prompt naming this file. Single-context leaf review of exact diff scope: no agents, nested reviews, or edits.

• fable-mode

  Hard multi-step work, uncertain diagnoses, repeated failures, 'did it work/is it fixed/prove it', or tasks needing careful verification + handoff. Skip routine changes.

• forge-repo-ui-skill

  Synthesize a repo-specific UI or design skill from current agent skills. Not ordinary UI work or backend-only work.

• handoff-audit

  Draft self-contained audit prompt for a separate task or reviewer: exact scope, falsifiable checks. Doesn't run the audit.

• impartial-review

  Independent review of code, a diff, or recent changes. Needs fresh reviewer context: exposed agents or authed Codex CLI.

• init-project

  Configure a starter project on setup request: detected project facts, only necessary user questions.

• lab

  Explicit ask to lab or prototype a visual, UI, motion, or game-feel element with live tuning before production.

• long-horizon

  Work spanning context windows or sessions, progress lost after compaction or retries, or 'work in verified rounds'.

• merge

  $merge only: self-review, Opus loop, CI, squash-merge.

• motion-design

  Animation that must read + feel right: explainers, promos, trailers, hero/UI motion, landing-page effects, generated scroll films, 3D embeds, component transitions + gestures, UI card loops, kinetic type, WebGL particles, janky animation. Not static layout.

• optimize-context

  Cut per-turn context or token load: trim kernels, skills, connectors; push generic context optimization to starter.

• opus-fullreview

  Full Claude review from Codex: Claude CLI runs impartial-review as Manager w/ fresh Opus sub-reviewers; Codex verifies each finding.

• perf-loop

  Measured optimization rounds with independent review: FPS, loading, latency, throughput, resources; no named target → triage every dimension first. Broad perf requests; skip isolated fixes.

• product-walkthrough

  Product tour or onboarding tour replaying real screens (product CSS + markup), guide cursor acts; embed or MP4.

• recall

  Before unfamiliar project-area work; retrieve project decisions or pitfalls; save authorized durable fact; quirk just cost retry, backed-out change, or user correction → pitfalls.

• redesign-concepts

  $redesign-concepts: built UI screenshots → image_gen → 10 redesigns → owner picks → build plan. Brief only → design-prototypes.

• refine

  Explicit workflow-improvement review; user preferences → rules or skill; recurring task friction justifying a narrow skill or reference change; unattended weekly review.

• security-loop

  Whole-repo security audit or pen-test in verified runs → fix owner's picks.

• seo-loop

  SEO/AEO/GEO: study the site, map how people search for it, compare with its field, improve pages. Use: 'improve SEO', keywords.

• servers

  After starting a dev server, or user asks what's running, which port is which, why browsers or node.exe are open, or to close old servers and browsers.

• session-hub

  Coordinate parallel agent sessions via shared append-only HTML hub. Use: 'run parallel sessions on this', joining a hub, or proactively when the checkout shows edits or commits this session didn't make.

• showpiece

  Push an artifact past what is expected of its kind, any medium. Use: ambitious creative direction, portfolio-quality work, replacing generic AI styling. Not quiet or faithful work (forms, dashboards, exact recreations, brand matching) → ui-quality or project UI skill. Skip routine edits unless invok

• sync-starter

  Pull template improvements into a spawned repo, compare starter drift, or push a generic improvement back to the starter.

• ui-quality

  UI quality floor (forms, dashboards, settings): build every state; audit by measuring the page (axe-core, focus, reflow, contrast). Use: 'audit accessibility', 'is this UI production-ready', 'check this page on mobile', 'feels wrong on my phone', 'break this UI' with worst-case data. Ambitious look

• untangle

  Whole-codebase structure survey: hidden links, needless links, surplus, special cases → decision log of fixes, each over the problems it solves → owner picks → $long-horizon fixes. Report only. Use: $untangle [area], 'untangle this codebase', 'where is this codebase tangled'. One diff → $impartial-r

• why

  Fresh critical second look at a recommendation: on $why, or before presenting own weighty pick (hard to undo, 2+ real options, real time or money at stake). Never from ordinary why questions.

• wow-animation

  $wow-animation <subject> [notes], or start a new shareable promo, explainer or trailer animation. Kickoff + orchestration; craft → motion-design.

• wow-loop

  Wow factor, target score ('get it to 8/10'), or visual work needing reference fidelity or repeated correction, one deliverable or set of like items. Skip routine cosmetic edits + talk about the skill.

• wrapup

  User asks if session is done, OK to archive, or what's left: one verdict from git, the PR, scratch files, running servers or browsers.

• writing-plans

  Clear task needing a multi-step implementation plan, dependency ordering, or durable handoff. Skip routine changes doable directly.

• writing

  Write or edit audience-facing prose: docs, UI copy, emails, release notes, application answers; voice matching, clarity review, AI-tell cleanup. Session replies follow session conventions.

• astra-fullreview

  Full multi-agent Codex review on gpt-6-astra, medium. Same lifecycle as codex-fullreview. Use: /astra-fullreview, 'full Astra review with sub-reviewers'.

• codex-fullreview

  Full multi-agent Codex review: codex exec runs impartial-review as Manager with fresh sub-reviewers (gpt-6.1-sol, high), then each finding verified. Use: /codex-fullreview, 'full Codex review with sub-reviewers'.

• codex-image-gen

  Image asset needed (icon, sprite, texture, splash, logo, marketing art, illustration) → Codex image_gen via codex exec. Use: 'generate an image', 'make an icon', 'use gpt-image / my Codex sub', placeholder → real art.

• coordinate

  Coordinate with another named Claude Code session: read its work + recs, critique, send own rec, let it answer, once agreed have it finish; ping user only when a step needs them. Use: /coordinate <session>, 'coordinate / review / go back and forth with another session'.

• gauntlet

  Only on explicit /gauntlet: draft → dare → final draft → why → rewrite → arena → commit, stop for /merge. Skills fully; other artifacts need declared apply method + real check.

• long-horizon-workflows

MCP servers (1)

• playwright-iso

  runs: npx @playwright/mcp@latest --isolated

Similar plugins