Emasoft/emasoft-plugins

Universal terminal-menu system for Claude Code plugins. Any agent invokes scripts/menu_write.py via Bash (~25ms) to render a JSON menu spec to a tempfile; a bundled Stop hook emits it as systemMessage at main-session turn end — the menu appears in the user terminal exactly when they can type a reply. No subagent fork, no extended-thinking overhead, no token cost to the orchestrator.

CLI plugin for installing, validating, auditing, searching, and managing Claude Code plugins from local sources or GitHub marketplaces. Includes 14 slash commands, an autonomous agent, CI/CD validation, Claude CLI validation, security audit, and 30+ checks.

Comprehensive validation suite for Claude Code plugins, marketplaces, hooks, skills, and MCP servers. Includes 17 Python validators, 2 expert agents, and 190+ validation rules.

Four-phase PR review pipeline and full codebase audit pipeline for Claude Code. PR review: code correctness swarm, claim verification, skeptical external review, security analysis with deduplication. Codebase audit: file inventory, grep triage, parallel discovery swarm, verification, gap-fill, consolidation, TODO generation, and optional fix loop.

Design documents, requirements analysis, architecture decisions - creates specifications for orchestrator

User's right hand - sole interlocutor with user, directs other roles (architect, orchestrator, integrator)

Exports current session segment (since last compaction) with system-reminder stripping -- main conversation, subagent transcripts, sidechains, and debug logs in structured markdown.

Multi-agent lifecycle management - spawns, monitors, hibernates, and terminates Claude Code agents across projects

Quality gates, testing, merging, release candidates - ensures code quality before integration

Task distribution, agent coordination, progress monitoring - executes architect's plans via subagents

General-purpose programmer agent for code implementation, testing, and debugging across multiple languages (Python, JS/TS, Rust, Go, .NET, C/C++, Swift).

Cross-platform clipboard operations for Claude Code agents. Native macOS/Windows/WSL support plus Linux scripts with Wayland, X11, and history tracking.

MCP server that offloads bounded LLM tasks from Claude Code to cheaper local (LM Studio, Ollama, vLLM, llama.cpp) or remote (OpenRouter) models. Profile-based configuration with ensemble mode, auto-batching, and secret scanning.

Run linters, formatters, and type-checkers without installing them using uvx, bunx, npx, and other no-install runners. Includes smart executor script and expert agent.

High-accuracy skill activation (88%+) with AI-analyzed keywords, weighted scoring, synonym expansion, and three-tier confidence routing.

Automatically reviews and fixes Claude's code changes using a separate Claude instance in a git worktree after commits. Two-phase pipeline: code correctness (Phase 1) then functionality verification (Phase 2).

Per-operation token usage reporter for Claude Code. Shows token counts, cost estimates, tool attribution, cache efficiency, duration, bash commands, web fetches, and file activity when agents complete. Only outputs in debug mode (claude --debug).