By cinjoff
Firehorse — lightweight cross-provider agentic skills framework, Claude-adapted distribution.
Analyzes requirements and codebase, generates context and meta-prompt
Lightweight subagent for direct delegated tasks
High-context decision-consistency oracle that protects inherited state and prevents drift
Creates implementation plans from context and requirements
Autonomous web researcher that searches, evaluates, and synthesizes a focused research brief
Run once after installing Firehorse for Claude Code. Checks Firehorse setup, detects Superset, and configures the user-scoped Superset MCP server safely. Use --check for read-only status.
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.
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
Triage issues through a state machine driven by triage roles. Use when user wants to create an issue, triage issues, review incoming bugs or feature requests, prepare issues for an AFK agent, or manage issue workflow.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Firehorse is a one-install agent skills distribution plus a small cross-provider foundation library.
In plain language: it collects the agent skills, subagents, web tools, code navigation tools, and setup glue that I want available in every serious coding agent session, then packages them so they work in both Pi.dev and Claude Code.
It exists because useful agent workflows are currently fragmented: some skills live in Claude plugin format, some tools are Pi extensions, some capabilities come from MCP, and some projects need orchestration context from Superset or other shells. Users should not have to install, wire, and audit each piece manually.
Firehorse gives that surface a reviewed, versioned home.
Firehorse is designed to support Pi.dev and Claude Code equally well. The recommended daily driver is:
Pi.dev + Codex model + firehorse-pi
Why this is the recommended path:
Claude Code remains first-class: the same selected upstream skills and shared subagent roles are mirrored into a Claude plugin. Claude just cannot consume Pi extensions directly, so the Pi distribution is the more complete driver today.
pi install npm:firehorse-pi
Or install this repository directly:
pi install git:github.com/cinjoff/firehorse
Then run:
/skill:firehorse-setup
/plugin marketplace add cinjoff/firehorse
/plugin install firehorse@firehorse
pnpm add firehorse
The core library is for adapter contracts and future integrations. It is not a skill runtime yet.
firehorse — TypeScript core library for provider/orchestrator contracts
and pinned upstream provenance.firehorse-pi — Pi.dev package that bundles selected upstream Pi packages
and Firehorse skills. This is the recommended user install.firehorse-claude — Claude Code plugin with mirrored skills and agents.Firehorse v0.1.0 is intentionally a foundation release. It does not include a Firehorse-authored skill runtime, prompt loader, slash-command runtime, or provider API transport yet.
┌───────────────────────┐
│ upstream sources │
│ skills, agents, tools │
└───────────┬───────────┘
│ reviewed + pinned
▼
┌─────────────────────────────────────────────────────────────────────────┐
│ firehorse │
│ │
│ ┌───────────────────────┐ mirrors / manifests ┌─────────┐ │
│ │ firehorse-core │────────────────────────────────▶│ Claude │ │
│ │ │ │ plugin │ │
│ │ provider contracts: │ └─────────┘ │
│ │ Claude / Codex / Pi │ ▲ │
│ │ │ │ │
│ │ orchestrators: │ install │
│ │ Superset / tmux / │ │ │
│ │ Conductor / terminal │ │ │
│ └───────────┬───────────┘ │ │
│ │ │ │
│ │ curated Pi allow-list │ │
│ ▼ │ │
│ ┌───────────────────────┐ recommended driver │ │
│ │ firehorse-pi │───────────────────────────────────────┘ │
│ │ │ │
│ │ Pi extensions │ context-mode, pi-lens, pi-mcp-adapter, │
│ │ Pi skills │ pi-mermaid, pi-subagents, pi-web-access, │
│ │ prompts + themes │ mattpocock/skills, impeccable │
│ └───────────────────────┘ │
└─────────────────────────────────────────────────────────────────────────┘
The important design choice: bundling is not the same as exposing.
firehorse-pi may bundle an upstream package for runtime support, but the Pi
manifest explicitly allow-lists what users see.
npx claudepluginhub cinjoff/firehorse --plugin firehorsefhhs-skills: All-in-one workflow plugin — 33 skills and 31 agents for spec-driven development, TDD, design quality, web research, startup validation, and autonomous execution. No other plugins required.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.
Memory compression system for Claude Code - persist context across sessions
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Editorial "Web Designer" bundle for Claude Code from Antigravity Awesome Skills.