human-frontier-labs-marketplace

Peer discovery and messaging between Claude Code sessions. UCAN-authenticated, multi-machine, push-delivered. Sessions see each other, send messages, share context automatically. Auto-installs the correct platform binary from GitHub releases on first use.

1 skill · 5 commands · 1 MCP server · hooks

Idiomatic Go architecture and design. Teaches Claude the five design instincts that separate Go code from translated Java/Python: zero-value construction, consumer-defined interfaces, embedding over inheritance, typed constants, and stdlib-first development. Proven: 10% to 60% interface usage rate, non-coder with skill beats CS-dev without.

1 skill · 1 agent

Set up and use the OpenCode 'zen' model harness in any project — route agent work to cheaper models by capability (fast=kimi-k2.6, reasoning-heavy/code-generation=glm-5.2) to save Claude's token budget, with Claude orchestrating and validating. One-command setup writes ~/.config/vinay-agent/config.toml and tests both models end-to-end; ask.sh calls a profile; /opencode-harness command included. Built by dogfooding — measured routing findings included. Use when integrating OpenCode, offloading reasoning/codegen to a cheap endpoint, or deciding which model a task should go to.

1 skill · 1 command

Automates Bubble Tea TUI design by analyzing requirements, mapping to appropriate components from the Charmbracelet ecosystem, generating component architecture, and creating implementation workflows. Use when designing terminal UIs, planning Bubble Tea applications, selecting components, or needing design guidance for TUI development.

1 skill

Expert Bubble Tea maintenance and debugging agent specializing in diagnosing issues, applying best practices, and enhancing existing Go/Bubble Tea TUI applications. Helps developers maintain, debug, and improve their terminal user interfaces built with the Bubble Tea framework.

1 skill

Meta-skill for building Claude skills that measurably improve output. Gather canonical docs (pith) -> detect where the model drops the ball one-shot AND multi-turn -> distill only the gap -> validate with a blind 4-cell benchmark and a multi-turn drift test. Use when creating/improving a skill, turning docs into a skill, or deciding whether a skill is worth building. Proven: a skill prevented multi-turn drift 13-0 across two stacks.

1 skill

Comprehensive coding standards and architectural patterns for the CareBridge eldercare management application. Includes critical patterns for case context management, Stripe integration, Next.js 15 requirements, component creation rules, and database operations.

1 skill

Chuqlab brand CSS design system for Tailwind CSS v4. Includes brand color palette (primary navy #141246, secondary red #EC1848, accent blue #0842EA), Geist font configuration, utility classes, and design guidelines with WCAG compliance.

1 skill

Build the Chuqlab Intelligence Engine (CIE) - a multi-tenant correctional intelligence platform. Guides phased, TDD-driven development of a GovTech law enforcement platform on GCP with Go backend, React/Next.js frontend, Whisper transcription, Ruvector vector search, and full TX-RAMP/CJIS compliance.

1 skill

Orchestrate gummy-agents across distributed network using 'dw' command for load-balanced, multi-host AI development. Automates distributed Claude Code agents for parallel development workflows.

1 skill

Design thinking for building systems with Claude Code. Covers hooks, daemons, agent pipelines, skills, multi-agent systems, and autonomous operations. Synthesized from Anthropic's agent research, AutoAgent, Boris Cherny's workflow, and production experience. The core finding: the bottleneck is harness design, not model capability.

1 skill

Embedded systems design for ESP32/Arduino. Built from a measured baseline: skips re-teaching the architecture Claude already does well and fills the two gaps it misses -- exact ISR primitives (IRAM_ATTR + FromISR + portYIELD_FROM_ISR) and measure-before-tune validation. Proven: blind 12-iter benchmark vs authority rubric, 24.25/25, beat no-skill baseline 12-0-0.

1 skill

Architecture guidance for cross-platform Expo apps (iOS, Android, web) from one codebase — keeps a growing app factored without over-building it. Share the brains (TypeScript logic), not the UI; add structure only when it earns its place. Covers expo-router, monorepo layout, platform divergence, NativeWind, EAS Build/Update, and version-specific SDK/config gotchas (autolinking, New Architecture, runtimeVersion, EXPO_PUBLIC secrets). Use when building or extending an Expo / React Native app. Proven: beat no-skill 9-1 pairwise on a multi-turn cross-platform drift benchmark.

1 skill

Operate tmux as an agentic multiplexer on a live, shared machine — run/coordinate parallel background work without mangling the user's or other agents' sessions. For Claude DRIVING the tmux CLI (not configuring tmux). Deterministic ownership (cc- namespacing, exact =name/%id targeting), native coordination (wait-for zero-poll, pane @options IPC), and lifecycle (track/teardown, never kill-server). Includes cc-tmux.sh lifecycle helper. Built measured-first: fills the gap (ownership + coordination primitives) and skips what the model already does well. Proven: closes the gap 3/3 vs no-skill, 6/6 safe on adversarial traps.

1 skill

Animated isometric architecture diagrams of your codebase from a small JSON spec: one request flow (network + sequence view, packets animated hop by hop) or a before/after of who uses whose code. Zero-dependency Node generator owns layout and lints the spec (no hand-placed coordinates, every hop and edge traced to file:line), outputs one self-contained HTML page with play/step, light/dark, phone layout. Icons: Isoflow isopack (MIT). Skill: /make-diagram.

1 skill

Manages distributed workloads and file sharing across Tailscale SSH-connected machines. Automates remote command execution, intelligent load balancing, file synchronization workflows, host health monitoring, and multi-machine orchestration using sshsync.

1 skill

Generate pictorial (breadboard-style) wiring diagrams for hardware builds from a small JSON spec — real Fritzing part photos + clean colored point-to-point wires, rendered offline and headless via Chromium. Use when asked to make, update, or fix a wiring diagram, hookup guide, or 'how do I wire this' picture for an ESP32/Pi/Arduino/discrete-component build. Pictorial, not schematic (the reader can't read schematics); not for PCB layout. v2: anti-convergence router with a measurable `score` subcommand, 65-part catalog (Wokwi auto-generation + Fritzing sourcing), lint_parts.py integrity gate (artwork aliasing, duplicate parts, provenance), 21-case adversarial test harness, circuit importers for Falstad and Wokwi files.

1 skill