Core MCP usage skill for Kusillo. Use when you need to initialize, discover, and call tools through POST /mcp with JSON-RPC 2.0.
**DEFAULT for MCP design tasks — dispatches reviewer/architect with MCP focus.**
Generate AI-written documentation for a specific directory or all undocumented directories. Creates structured READMEs optimized for both human and AI consumption, and updates the…
Scan project and report documentation coverage. Finds undocumented directories, stale manifest entries, and prioritized recommendations.
Use when executing a multi-step workflow that crosses MCP servers (e.g., GitHub PR → memory write, Calendar event → fetch context → prep brief).
Design and implement MCP gateway behavior for paid contract tools. Use when mapping ABI to tools, defining handlers, and orchestrating paid invocation pipelines.
Internal guardrail invoked by mcp-builder only. Audits MCP tool definitions that contain list_*, search_*, find_*, get_all_*, or SQL-executing operations for missing…
Audit MCP server registrations across user, project, and plugin configs — report always-on servers, duplicates, and disable recommendations
Design an MCP server's tool surface by analyzing an API and proposing interaction categories. Produces a plan consumable by /mob:swarm for implementation.
Convert legacy subagents, slash commands, and rules to the unified Agent Skills format. Use when migrating agent resources, consolidating legacy directories, or standardi — from…
Use when auditing or setting up a repository for multi-agent development across Claude Code, Cursor, and Codex — or making a repo compatible with those assistants.
マルチエージェント環境（Claude Code / Codex / GitHub…
Use when the user says "save state", "/mxSave", "session end", "before /compact", "wrap up", or otherwise wants to persist the current mx-project state (clean settings, update…
Autonomous trading agent for Nad.fun that scans markets, analyzes tokens, executes trades, and distributes profits to MMIND token holders.
One repeatable pass that keeps agent-steer prose leading with the AFFORDANCE, not the prohibition. Runs `fak score negframe` over the steer-prose corpus (AGENTS.md, CLAUDE.md, the…
Orchestrates multi-agent task execution using a Royal Navy squadron metaphor — from mission planning through parallel work coordination to stand-down.
Run the full agent-team pipeline end-to-end for a single feature — branch + worktree, PM grooms a Tasks Plan, human approves the plan, inner SWE/Tester loop per task, PM…
Assess patent novelty and non-obviousness against prior art. Use when user says \"专利查新\", \"patent novelty\", \"可专利性评估\", \"patentability check\", or wants to evaluate if — from…
Get an external patent examiner review of a patent application. Use when user says \"专利审查\", \"patent review\", \"审查意见\", \"examiner review\", or wants critical feedback — from…
Plan Mode の `
` を出す直前に、計画の段階・依存関係・検証観点を一時HTMLで視覚確認したい時に使う agent-internal skill。Use this agent-internal skill to create a temporary HTML preview for a plan just…
A focused Cloud Run Admin API CLI with agent-native output, local inventory, and command discovery. Trigger phrases: `inspect Cloud Run services`, `list Cloud Run jobs`, `find…
Multi-agent PR review — component, integration, and routing analysis merged into a structured audit doc with severity and PR buckets.
Rigorous mathematical proof verification and fixing workflow. Reads a LaTeX proof, identifies gaps via cross-model review (Codex GPT-5.4 xhigh), fixes each gap with full — from…
Enable cross-machine Psyche context sync via a private GitHub repo (spt-agent-storage). Drives install + auth checks for the `gh` CLI, creates the repo via `gh repo create…
NVIDIA RAG Blueprint — deploy, configure, troubleshoot, and manage. Handles any RAG action: deploy, install, start, enable, disable, toggle, change, configure, troubleshoot,…
Use when the user wants to create a new GitHub repository from scratch with production-ready infrastructure.
Use when auditing or polishing a repository to Tamir Cohen standards — README badges (author, version, AI targets), Makefile install/update/uninstall, docs tree, CI/CD, changelog…
Audit and fix CLAUDE.md files using a tiered binary checklist based on official Anthropic best practices and community guidelines.
Analyze a feature or sprint for scope creep by comparing current scope against the original plan. Flags additions, quantifies bloat, and recommends cuts.
Langfuse production trace를 관찰하고 품질 regression·비용 이상·실패 패턴을 자동 분석한 뒤 prompt·skill 수정안을 PR draft 형태로 제안한다.
Search published venue papers (IEEE, ACM, Springer, etc.) via Semantic Scholar API. Complements /arxiv (preprints) with citation counts, venue metadata, and TLDR.
On-demand parent-supervised delegation of a single real task to Claude TUI as a subagent — briefings, checkpoints, failure escalation, and completion evidence.
On-demand parent-supervised delegation of a single real task to Codex TUI as a subagent — briefings, checkpoints, failure escalation, and completion evidence.
On-demand parent-supervised delegation of a single real task to Cursor agent TUI (cursor-agent) as a subagent — briefings, checkpoints, failure escalation, and completion…
On-demand parent-supervised delegation of a single real task to OpenCode CLI as a subagent — briefings, checkpoints, failure escalation, and completion evidence.
On-demand parent-supervised delegation of a single real task to Pi CLI as a subagent — briefings, checkpoints, failure escalation, and completion evidence.
Reverse engineer and analyze a specific feature in Claude Code's source. Use when the user wants to understand how a Claude Code feature works internally, or when asked to analyze…
Creates Claude subagents equipped with curated skills for solving problems within DAG workflows. Use when designing a specialist subagent, selecting skills for a subagent, writing…
Write, review, debug, and explain SLANG flows. Use when creating multi-agent workflows, writing .slang files, designing agent orchestration, or when the user asks about SLANG…
Vocalize Claude's last response using local Qwen3-TTS. Default voice is the Oracle (deep, resonant Dune narrator). Use --preset for emotion-controlled preset speakers.
Scaffold a new DovePaw background agent end-to-end. Creates agent files in ~/.dovepaw/tmp/ so the agent appears immediately in the Kiln sidebar group, ready to test.
Cross-repo migration swarm — one coordinator + N parallel subagents (one per target repo) that apply the same transformation, open PRs, wait for CI, and report back to a shared…
Use when switching between Claude Code, Cursor, and Codex mid-task without losing context — hand off work via GitHub issue Resume blocks, resume on another platform, or list…
Multi-agent debate system supporting 6 AI providers (Gemini, OpenAI, DeepSeek, Groq, Grok, Mistral)
Deep collaborative analysis pipeline. Multi-perspective exploration, deep analysis, user-driven discussion loops, and cross-perspective synthesis.
Use the Telarchy API at https://telarchy.com/api. Telarchy is the approval layer for actions, for any agent, human or AI: the owner defines the metrics they value, participants…
Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.
Execute tasks through systematic exploration, pruning, and expansion using Tree of Thoughts methodology with meta-judge evaluation specifications and multi-agent evaluation
Diagnosis and repair guide for errors and failures including CI. Use when user mentions something broken, errors, it doesn't work, CI failures, CIが落ちた, build errors, test — from…
Diagnosis and repair guide for errors and failures including CI. Use when user mentions something broken, errors, it doesn't work, CI failures, CIが落ちた, build errors, test — from…
Visually verify Ghostfolio UI state using agent-browser after agent actions. Use after /uat add_activity calls to confirm the UI reflects the change.
Comprehensive MCP server management. Discovers, updates, and helps set up MCP servers with dependency detection and best practice guidance.
official-llms-txts 配下の公式ドキュメント (llms.txt / llms-full.txt) の更新差分を、人間向けの changelog / リリースノート風 Markdown として生成する。対象サイトは --site で切り替える (claude-code-docs / mcp)。詳細版を LLM…
Use when the user wants to refresh the RTK savings stats displayed on the claude-code-codex homepage.
Use when contributing to the Valkey server source - C internals, event loop, data structures, cluster, replication, persistence, memory, threading, modules, scripting, tests,…
Manage MCP servers - discover, analyze, execute tools/prompts/resources. Use for MCP integrations, intelligent tool selection, multi-server management, context-efficient…
Make a web app agent-ready — propose a WebMCP tool manifest, integrate, verify in a real browser, heal; unrelated code stays untouched.
Run multi-agent simulations to measure output consistency. Use when you want to A/B test, validate behavioral equivalence, or stress-test non-deterministic behavior at scale.
Write agent-consumed markdown well at the moment of writing — CLAUDE.md or AGENTS.md content, .claude/rules files, agent-loaded reference/context docs, navigation-pointer lines,…
Project-aware multi-agent code review for Hurrah.tv. Asks for a mode (quick / review), runs reviewers, fixes score-50+ findings inline by default, and always surfaces follow-up…