Menu
SkillsMP has collected 161 skills from DreamLab-AI/agentbox. Open a skill to review its source and details.
Showing 40 of 161 collected skills.
Implement features with tests and quality gates, debug hard bugs, and stress-test designs. Use when building a feature with TDD/EDD, chasing a stubborn multi-function bug (feedback-loop-first), interrogating a design before coding, or running a…
Plan-execute-reflect loop over a persistent Python kernel (the `code-interpreter` MCP): write Python, execute, observe the returned trace (stdout / exception / last value), then revise. Variables, imports, and dataframes survive across every tool call in the…
Fires as a post-task hook after a non-trivial completed task (3+ tool calls, observable terminal outcome) to distil 0-N generalisable IF/THEN lessons from the trajectory and store them in RuVector so they surface at the start of similar future tasks; also…
Unified dispatcher for the full skills estate. Use when you don't know which skill to invoke — describe your task and get routed to the optimal skill, agent composition, or MCP tool. Not for executing the task itself: it only routes, then hands off to the…
Execution-gated branching code generation (ADR-020 Surface 2). Generate N candidate solutions (default ≤5) by invoking `sparc:coder` with varied temperature/framing, execute each in a fresh `KernelSession` via the `code-interpreter` (ADR-018 kernel) MCP,…
Store and retrieve verified, executable Python skill primitives (a function plus assertions plus at least one example) in the code-harness procedural memory tier. Use when a reusable utility, parser, validator, or algorithm has proven general-purpose across…
Trigger on "/podcast-ingest", "weekly podcast ingest", "process new podcast episodes into the ontology", or setting up/debugging the podcast-cron schedule. Weekly cron that downloads new episodes from configured YouTube podcasts, extracts evidence-backed…
Audit and de-slop prose for AI writing tells (lexical, structural, and narrative), strip technical AI provenance marks (invisible Unicode watermarks, statistical sampling fingerprints, C2PA/EXIF/XMP metadata), and enforce UK English. Use when writing or…
Implement persistent memory patterns for AI agents using AgentDB — session memory, long-term storage, pattern learning, hierarchical memory, consolidation, ReasoningBank migration. Use when building stateful agents, chat systems, or intelligent assistants.…
Use when building RAG pipelines, running semantic/similarity vector search, optimising search speed, tuning HNSW indexing or quantization, or scaling to millions of vectors with local AgentDB. Covers CLI, TypeScript API, caching, batch operations, and MCP…
Queen-led multi-agent coordination in Claude Flow — one strategic queen directs specialized workers through structured voting and shared persistent memory. Use when a task needs formal consensus (majority, weighted, or Byzantine fault tolerance) and…
Diagrams-as-code routing hub. Two engines: (1) diagram-design — editorial-quality self-contained HTML/SVG diagrams with 28 visual types, branded design system, semantic patterns, draw.io/Mermaid import, dark/light/terminal variants, animation, and accessible…
Generate publication-quality LaTeX reports — white papers, sector analyses, policy briefs, technical documentation — with data-driven charts, diagrams-as-code, citations, and multi-agent quality review. Use when a document needs professional typesetting with…
Architect's playbook and capability catalog for the RuVector monorepo (Rust crates, npm packages, WASM builds). Use when a task could be served by a RuVector capability — vector/hybrid search, graph intelligence, self-learning embeddings, attention…
Search and answer questions about the owner's private personal email archive via the local Private Email MCP Gateway. Use when the user asks anything that needs looking in their personal mail — "did I get an invoice from X", "when did Y email me about Z",…
Advanced AgentDB beyond single-database vector search: distributed QUIC sync across nodes, multi-database coordination and sharding, custom distance metrics, hybrid vector+metadata search, MMR diversity, context synthesis, and reinforcement-learning plugins.…
Complete visual content system for Claude Code. Use when generating blog headers, infographics, editorial art, diagrams, or comics. Default aesthetic: light backgrounds, teal/burnt orange accents, hand-drawn sketch style.
Consolidated creative UI/UX skill combining design vision and production-grade implementation. Two modes: --design (ask-first, bold creative direction) and --build (production frontend code). Anti-AI-slop, distinctive aesthetics, shadcn/Tailwind/Phosphor…
DEPRECATED — replaced by bencium-creative. Use bencium-creative with --design flag for bold creative UX and --build flag for production frontend implementation.
DEPRECATED — replaced by bencium-creative. Use bencium-creative with --design flag for bold creative UX and --build flag for production frontend implementation.
AI-first development methodology with specification-driven artifact traceability. PRD → SPEC → ADR → TASK → CODE → REVIEW → DEPLOY pipeline with AI-native ADRs (model selection, prompt strategy, agent orchestration), LLM evaluation suites, guardrails…
Use when publishing a book, preparing an arXiv submission, formatting a manuscript for KDP, or turning a markdown manuscript into a print-ready PDF. End-to-end pipeline orchestrating parallel agent swarms for LaTeX (memoir) conversion, BibTeX citation…
Keyword web search that returns long page extracts for grounding an LLM in dense source context rather than a synthesized answer. Use when you need exact-match keyword retrieval over the live web — specific entities, dates, or locations — and want raw source…
OpenAI Codex / GPT-5.4 integration for Claude Code. Use when the user says "consult with openai", "talk to codex", "ask gpt-5.4", "get a second opinion from openai", "delegate to codex", "openai review", or "codex rescue". Provides code reviews, adversarial…
Generate AI images and video with ComfyUI's node-based workflows (FLUX, Stable Diffusion, video models) on a local GPU or distributed Salad Cloud compute. Use when a task needs text-to-image, image-to-image, or text/image-to-video generation, ComfyUI workflow…
Estimate GPU endpoint costs, agent job costs, and MRC20 token operations for the DreamLab AI ecosystem. Use when pricing inference/image-gen/analytics endpoints, calculating agent job holds and settlements, converting between DREAM tokens and sats, or valuing…
AI-powered CUDA development with 4 specialist agents (General, Optimizer, Debugger, Analyzer) plus an MCP toolset. Use when writing CUDA kernels (.cu/.cuh), optimising GPU code (coalescing, shared memory, occupancy), debugging nvcc compilation or race…
Use when the user says "ask deepseek", "consult deepseek", "delegate reasoning to deepseek", or needs advanced multi-step Chain-of-Thought reasoning, deep debugging root-cause analysis, or dependency-aware task planning via DeepSeek's special reasoning…
Control and inspect the nightly dream machine — the dream-engine loop that runs evidence-gated repository evolution overnight (ADR-052). Use when the user says "/dream", "dream machine", "dream engine", "did anything dream last night", asks about the dream…
Real-time AI meeting copilot. Dual audio capture (system + mic), live transcription (faster-whisper local or Deepgram cloud), LLM coaching loop (Claude/GPT every ~35s), always-on-top overlay, session logging with markdown export. Use when the user says…
Generate isometric network-topology, infrastructure, and architecture diagrams as FossFLOW JSON (compact LLM-optimised format or full verbose SVG/JSON with zones and labels). Trigger on "isometric diagram", "network topology diagram", "FossFLOW",…
Use when creating GitHub issues, managing project boards, planning sprints, decomposing work into vertical-slice tracer bullets, or triaging an issue backlog with swarm coordination. NOT for PR review (use github-code-review), release versioning (use…
Single-agent Godot 4 work: write/debug GDScript or C# scripts, edit scenes and node systems, wire signals, physics, navigation, shaders, export builds, godot-rust (gdext) native extensions, and OpenXR/WebXR integration. Use for targeted Godot scripting…
Schedule recurring agent tasks on cron/interval/one-shot schedules — start with /hermes-scheduler. Triggers: 'run this every 30m', 'schedule a daily 9am job', 'cron job that invokes Claude Code', 'background scheduler for agent tasks'. A standalone Python…
Trigger on "https bridge", "CORS error accessing host", "debug host webserver", "reach host dev server from container", or needing a secure-context (HTTPS) origin for a host HTTP server (Vite/Next/Express on the Docker host). Stands up a self-signed HTTPS…
Context, persistence, and process-tracking companion for single-agent Claude Code sessions. Use when hydrating session context (git/plan/memory), tracking phased read/plan/implement/validate/document task plans, running blueprint workflows, doing progressive…
Trigger when creating Google NotebookLM notebooks, ingesting sources (URLs, PDFs, YouTube, Drive, text), chatting with those sources, or generating NotebookLM artifacts — audio overviews (podcasts), video explainers, slide decks, quizzes, mind maps, reports,…
Generate brand-constrained, quality-gated visual artifacts instead of freestyle output. Use when the user asks for a web prototype, landing/marketing page, SaaS dashboard, mobile mockup, slide deck, email template, or any HTML visual artifact that must look…
Delegate a coding or reasoning task to OpenAI Codex (GPT-5.4) via MCP for a second opinion from a non-Claude model. Use when you want to cross-check a hard algorithm, design, or refactor against a different frontier model, or get an independent…
Generate publication-quality academic diagrams and statistical plots from text via a multi-agent VLM pipeline with iterative refinement (OpenAI or Google Gemini / nano-banana-2). Use when creating methodology diagrams, architecture overviews, flow charts,…