emasoft-plugins

High-accuracy skill activation (88%+) with Dewey-like domain classification, dynamic tool catalog from skill index, weighted scoring, and three-tier confidence routing.

Comprehensive validation, management, and standardization suite for Claude Code plugins and marketplaces. Includes 190+ validation rules, plugin lifecycle management, marketplace operations, health checks, security auditing, GitHub repo validation, plugin/marketplace repo scaffolding, and standardization tooling. Features severity hierarchy, --strict mode, language-aware token estimation, and universal plugin/marketplace templates.

Design documents, requirements analysis, architecture decisions - creates specifications. Requires AI Maestro for inter-agent messaging.

User's right hand - sole interlocutor with user, directs other roles. Requires AI Maestro for inter-agent messaging.

Quality gates, testing, merging, release candidates - ensures code quality. Requires AI Maestro for inter-agent messaging.

Task distribution, agent coordination, progress monitoring - executes plans via subagents. Requires AI Maestro for inter-agent messaging.

Multi-project agent management - staff planning, lifecycle, approval workflows, failure recovery. Requires AI Maestro for inter-agent messaging.

General-purpose programmer agent that executes tasks assigned by the Orchestrator. Uses globally installed skills and SERENA MCP for code navigation.

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

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.

Ultracode code-audit engine for Claude Code: one deterministic map-filter-reduce workflow (scripts/workflows/caa-engine.js) drives every operation over swarms of opus agents — pre-commit gate of staged files, whole-codebase or scoped scan, delta recheck since a git ref, scan-and-fix with per-fix adversarial verification, and GitHub PR review with claim-verification and cross-layer lenses plus 21 distilled domain lenses (docker, solidity, iOS, GraphQL, frontend, JWT, prompt-injection, ...). Every finding is adversarially verified before it is reported; one consolidated report plus machine-readable findings.json per run.

Per-operation token usage reporter for Claude Code 2.1.108+. Shows token counts, cost estimates, tool + skill + agent attribution, cache invalidation detection, worktree sub-agent breakdown, compact_boundary markers, CLAUDE.md reload events, and file activity when agents complete. Only outputs in debug mode (claude --debug).

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.

Prompts Claude to review and fix code after git commits using a blocking /recheck skill

Universal terminal-menu system for Claude Code plugins. Any agent invokes scripts/menu_write.py via Bash 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. ~25ms, no subagent fork, no extended thinking overhead.