Use this skill to generate a Midnight Network DApp spec brief through an adaptive Q&A. The output is a concise brief (not a full spec) intended as input to a downstream spec / planning skill such as /superpowers:brainstorming or /sdd:specify. Triggers on "generate midnight dapp spec", "improve midnight dapp prompt", "midnight dapp brief", "midnight dapp prompt", or /prompt-rewriter:midnight-dapp-spec.

A fast, gitignore-aware command-line tool for recursively searching file contents by regular expression — a smarter, faster `grep`. Use when searching code or text for a pattern across a directory tree, filtering matches by file type or glob, listing/counting matching files, doing search-and-replace previews, or running gitignore-aware code search. Triggers on mentions of ripgrep, the `rg` command, "search the codebase for", "grep for", "find all files containing", or replacing `grep`/`ack`/`ag`. This is the rg CLI tool, NOT the fuzzy-search MCP server (which wraps rg).

Guides users through their first trade on Senpi/Hyperliquid. Walks through discovery (top traders), creating a mirror strategy with a chosen trader, monitoring, and closing the strategy. Use when user says "let's trade", "first trade", "teach me to trade", "how do I trade", or when state is AWAITING_FIRST_TRADE. Can also run when state is not READY (e.g. after entrypoint Step 3); then prompts for wallet funding before starting when needed. Requires Senpi MCP to be connected.

Audit Codex/OpenClaw skills: loaded roots, duplicate skills, unused skills, prompt-budget costs, compact descriptions.

Use PROACTIVELY when creating new Claude Code skills from scratch. Automated generation tool following Claudex marketplace standards with intelligent templates, pattern detection, and quality validation. Supports guided creation, quick start templates, clone-and-modify, and validation-only modes. Not for modifying existing skills or non-skill Claude Code configurations.

Use PROACTIVELY when users need guaranteed schema compliance or validated tool inputs from Anthropic's structured outputs feature. Expert advisor for choosing between JSON outputs (data extraction/formatting) and strict tool use (agentic workflows). Analyzes requirements, explains trade-offs, and delegates to specialized implementation skills. Not for simple text responses or unstructured outputs.

Manages parent/child agent relationships with task delegation and result aggregation. Supports sequential chains, parallel fans, conditional routing, retry logic, timeout handling, and YAML-based visual workflow definition.

Toolkit for comprehensive Spreadsheet reading, creation, editing, and analysis with visual quality control. Use to work with spreadsheets (.xlsx, .xlsm, .csv, .tsv) for: (1) Creating new spreadsheets with formulas and formatting, (2) Reading or analyzing tabular data, (3) Modifying existing spreadsheets while preserving formulas, (4) Building financial models with proper formatting, (5) Data visualization with in-sheet charts, or any other spreadsheet tasks.

Integrate LangChain applications with You.com tools (web search, content extraction, retrieval) in TypeScript or Python. Use when developer mentions LangChain, LangChain.js, LangChain Python, createAgent, initChatModel, DynamicStructuredTool, langchain-youdotcom, YouRetriever, YouSearchTool, YouContentsTool, or You.com integration with LangChain.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.

Orchestrate teams of parallel Claude Code sessions working on the same codebase. Handles task decomposition, agent coordination, context isolation, and merge strategies. Builds on worktree-manager for infrastructure.

Architect, execute, and iterate on AI evaluations using the Data-Task-Score framework. Treats evals as the modern, quantifiable version of a PRD. Use when the user asks to "build an eval," "improve model quality," "test an agent workflow," "quantify product intuition," "move beyond vibe checks," "measure AI output," "score LLM responses," "benchmark a prompt," or "set up evaluation infrastructure." Also triggers on phrases like "how do I know if this is working," "is the model getting better," or "eval-driven development."

Lead generation and prospecting using anysite MCP server for LinkedIn prospect discovery, email finding, company research, and contact enrichment. Extract contacts from websites, find decision-makers at target companies, and build qualified prospect lists for sales, recruiting, and business development. Supports LinkedIn (primary), web scraping for contact extraction, and Instagram business discovery. Use when users need to build prospect lists, find decision-makers, extract contact information, research potential customers, or enrich existing leads with additional data.

INVOKE THIS SKILL when auditing an AI agent or LLM app for regulatory compliance. Covers EU AI Act, GPAI Code of Practice, GDPR, NIST AI RMF, Colorado AI Act, HIPAA, and ISO 42001. Scans the codebase for compliance gaps, cross-references Arize instrumentation for audit trail coverage, and produces an actionable remediation checklist tailored to the selected frameworks.

Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, deploying persistent autonomous agents, or building complex multi-step AI automation systems.

Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, deploying persistent autonomous agents, or building complex multi-step AI automation systems.

Automate Chrome browser tasks using agent-browser CLI. Navigate pages, fill forms, click buttons, take screenshots, extract data, and replay recorded workflows — all inside the user's real Chrome session.

OpenAI Codex CLI configuration guide for Claude. Use when users ask about: codex config, ~/.codex/config.toml, codex sandbox policies, codex approval modes, creating codex skills, codex mcp servers, codex model settings, local LLM providers, or integrating tools with Codex.

Data format specialist covering TOON encoding, JSON/YAML optimization, serialization patterns, and data validation for modern applications. Use when optimizing data for LLM transmission, implementing high-performance serialization, validating data schemas, or converting between data formats.

Multiagent AI system for scientific research assistance that automates research workflows from data analysis to publication. This skill should be used when generating research ideas from datasets, developing research methodologies, executing computational experiments, performing literature searches, or generating publication-ready papers in LaTeX format. Supports end-to-end research pipelines with customizable agent orchestration.

Use when verifying outcomes with code instead of LLM judgment, versioning prompts with hashes, or ensuring reproducible agent behavior. Load for any critical verification. Scripts return boolean exit codes, not subjective assessments. Prompts use semantic versioning with SHA256 validation.

DSPy: declarative LM programs, auto-optimize prompts, RAG.

A fast, user-friendly command-line tool for finding files and directories by name — a simpler, faster `find` replacement. Use when searching for files or directories by name or regex/glob pattern, filtering results by type, extension, size, or modified time, respecting .gitignore, or running a command per result with -x/-X. Triggers on mentions of the fd command, fdfind, "find files named", "search for a file", or replacing `find`. This is the fd CLI tool, NOT the file-search MCP server.

Prompt engineering guidance for Gemini (Google) model. Use when crafting prompts for Gemini to leverage system instructions, multimodal capabilities, ultra-long context, and strong reasoning features.

Use when inspecting a ghcrawl SQLite store, pulling GitHub issue/PR data, refreshing summaries, embeddings, and clusters, or extracting one cluster and its evidence through the ghcrawl CLI.

Use when upgrading a Laravel 12 application to Laravel 13.0. Covers PHP 8.3 requirement, composer commands, breaking changes (PHPUnit 12, Pest 4, VerifyCsrfToken → PreventRequestForgery, cache prefixes, serializable_classes, pheanstalk 8.0+), Attributes migration (Eloquent + Queue), Symfony 7.4/8.0 upgrade, and Laravel Boost MCP automated upgrade.

Search Argentine legal databases (SAIJ, JUBA, CSJN, JUSCABA) for jurisprudence, legislation, case summaries, and doctrine using the `ley` CLI. Use when the user asks about Argentine law, court decisions, legal precedents, fallos, jurisprudencia, legislación, or mentions SAIJ, JUBA, CSJN, JUSCABA. Supports parallel search across databases, JSON/table/text output, and filtering by jurisdiction. Built from reverse-engineered MCP servers (hernan-cc) — direct HTTP calls, no MCP layer needed.

Execute programs on a compiled transformer stack machine where every instruction fetch and memory read is a parabolic attention head. Demonstrates that transformer attention + FF layers can implement a working computer. Use when user mentions "llm-as-computer", "lac", "stack machine", "compiled transformer", "percepta", "parabolic attention", "execute program", or asks to run/trace programs on the transformer executor.

This skill should be used when the user asks for "model council", "multi-model", "compare models", "ask multiple AIs", "consensus across models", "run on different models", or wants to get solutions from multiple AI providers (Claude, GPT, Gemini, Grok) and compare results. Orchestrates parallel execution across AI models/CLIs and synthesizes the best answer.

Agent continuity and cognitive health infrastructure. Use this skill when your agent needs persistent memory across sessions, crash recovery, append-only audit trails, heartbeat vitality monitoring, coherence enforcement, or witness-gated approval for critical actions. Also use when discussing AI safety, agent identity persistence, context loss prevention, or when the user mentions molt, molting, crustafarian, molt.church, or agent health. Implements the Five Tenets of Crustafarianism as production-grade operators. Built by X-Loop³ Labs (Switzerland) from 100k+ agents on Moltbook.

Help with MongoDB query optimization and indexing. Use only when the user asks for optimization or performance: "How do I optimize this query?", "How do I index this?", "Why is this query slow?", "Can you fix my slow queries?", "What are the slow queries on my cluster?", etc. Do not invoke for general MongoDB query writing unless user asks for performance or index help. Prefer indexing as optimization strategy. Use MongoDB MCP when available.

Help with MongoDB query optimization and indexing. Use only when the user asks for optimization or performance: "How do I optimize this query?", "How do I index this?", "Why is this query slow?", "Can you fix my slow queries?", "What are the slow queries on my cluster?", etc. Do not invoke for general MongoDB query writing unless user asks for performance or index help. Prefer indexing as optimization strategy. Use MongoDB MCP when available.

This skill should be used when the user asks to "set up Notion MCP", "explain OAuth flow", "understand Notion MCP", "how to authenticate with Notion", or wants to learn about the Notion MCP setup process. Provides understanding of what Notion MCP is and why OAuth-based setup is required.

Use when the user asks to create, scaffold, or edit Jupyter notebooks (`.ipynb`) for experiments, explorations, or tutorials; prefer the bundled templates and run the helper script `new_notebook.py` to generate a clean starting notebook. Originally from OpenAI's curated skills catalog.

Build search applications with OpenSearch from scratch. Use this skill when the user mentions search app, index setup, search architecture, semantic search, vector search, hybrid search, BM25, dense vector, sparse vector, agentic search, RAG, embeddings, KNN, PDF ingestion, document processing, or any related search topic. Activate even if the user says search quality, evaluation, nDCG, precision, relevance tuning, or search builder without mentioning OpenSearch.

Outlines: structured JSON/regex/Pydantic LLM generation.

Structured prompt construction and quality validation for Jerry Framework. Invoke when building structured prompts, generating NPT-009/NPT-013 constraints, or scoring prompt quality. Guides users through the 5-element prompt anatomy, generates formatted constraints with XML wrapping, and scores prompts against the 7-criterion rubric.

Evaluate, optimize, and enhance prompts using 58 proven prompting techniques. Use when user asks to improve, optimize, or analyze a prompt; when a prompt needs better clarity, specificity, or structure; or when generating prompt variations for different use cases. Covers quality assessment, targeted improvements, and automatic optimization across techniques like CoT, few-shot learning, role-play, and 50+ more.

Review a proposed Agent Skill for structural validity and content quality before publishing. Runs the skill-validator CLI to check for structural issues, scores the skill with an LLM judge, and interprets results to advise SMEs on what to address. Use when a user wants to review, validate, or quality-check an Agent Skill.

Execute pending tasks in dependency order with wave-based concurrent execution via Agent Teams

Skill that analyzes 18-month scenarios from a news headline. Runs the primary analysis with the scenario-analyst agent and obtains a second opinion with the strategy-reviewer agent. Generates a comprehensive English report covering 1st/2nd/3rd-order impacts, recommended stocks, and a critical review. Example: /scenario-analyzer "Fed raises rates by 50bp" Triggers: news analysis, scenario analysis, 18-month outlook, medium-to-long-term investment strategy

Build search applications with OpenSearch. Use this skill when the user mentions search app, index setup, search architecture, semantic search, vector search, hybrid search, BM25, dense vector, sparse vector, agentic search, RAG, embeddings, KNN, PDF ingestion, document processing, search quality evaluation, or any related search topic.

Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.

Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.

Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.

Guide for using SLIME (LLM post-training framework for RL Scaling). Use when working with SLIME for reinforcement learning training of language models, including setup, configuration, training execution, multi-turn interactions, custom reward models, tool calling scenarios, or troubleshooting SLIME workflows. Covers GRPO, GSPO, PPO, Reinforce++, multi-agent RL, VLM training, FSDP/Megatron backends, SGLang integration, dynamic sampling, and custom generation functions.

Automatically convert Confluence specification documents into structured Jira backlogs with Epics and implementation tickets. When an agent needs to: (1) Create Jira tickets from a Confluence page, (2) Generate a backlog from a specification, (3) Break down a spec into implementation tasks, or (4) Convert requirements into Jira issues. Handles reading Confluence pages, analyzing specifications, creating Epics with proper structure, and generating detailed implementation tickets linked to the Epic.

A skill that analyzes user requirements to generate system prompts ready for evaluation. It determines whether a single or multi-prompt architecture is needed and queries for missing information if requirements are insufficient.

Generate comprehensive, publication-quality technical manuals with thematic storytelling using multi-agent orchestration. Use when user asks for themed documentation, narrative technical guides, memorable training materials, or mentions themes like Victorian, steampunk, art deco, medieval, pharmaceutical, or nautical combined with technical topics. Triggers on phrases like "create themed manual", "documentation with storytelling", or requests to make docs "more engaging" or "memorable".

Convert a public webpage URL into Markdown and save it as a reusable `.md` file with the bundled script. Prefer `https://r.jina.ai/<url>` first, and only fallback to `https://markdown.new/` if `r.jina.ai` is unavailable. Use this whenever the user wants to turn a public webpage, article, documentation page, blog post, release note, or reference URL into Markdown for reading, archiving, summarizing, extraction, RAG prep, or downstream agent reuse, even if they do not explicitly mention markdown or saving a file.

Build and refactor CLIs for AI agent compatibility. Use when making command-line interfaces machine-readable, adding structured JSON output, hardening inputs against hallucinations, implementing safety rails like dry-run flags, adding schema introspection, or designing multi-surface architectures (CLI + MCP).

Create an AI Product Strategy Pack (thesis, prioritized use cases, system plan, eval + learning plan, agentic safety plan, roadmap). Use for AI product strategy, LLM/agent strategy, AI roadmap, AI-first product direction.

Scans project structure and agent system components, runs 11-category weakness analysis, and produces actionable restructuring recommendations. Use when analyzing an existing agent system for weaknesses. Use when user says "analyze agent system", "check agent setup", "audit agent config". Use when called by migrating-agent-systems or refactoring-skills.

INVOKE THIS SKILL for Arize Prompt Hub and `ax prompts` workflows: author or import templates and save (Workflows A–B), label/promote (C), or list/get/edit/delete/duplicate (D). Use when the user mentions ax prompts, Prompt Hub, creating/editing/saving a prompt, `{variable}` placeholders, or production/staging labels. For improving prompt text using traces or eval scores, use arize-prompt-optimization. For running experiments, use arize-experiment.

This skill should be used when the user asks to "model agent mental states", "implement BDI architecture", "create belief-desire-intention models", "transform RDF to beliefs", "build cognitive agent", or mentions BDI ontology, mental state modeling, rational agency, or neuro-symbolic AI integration.

Guides the agent through upgrading a Capacitor app project to a newer major version. Supports upgrades from Capacitor 4 through 8, including multi-version jumps. Performs all upgrade steps manually — does not use the interactive `npx cap migrate` command. Do not use for plugin library upgrade or non-Capacitor mobile frameworks.

Comprehensive cardiology editorial writing system for thought leadership newsletters. Use when the user wants to: (1) Identify and score recent landmark trials from top cardiology journals (NEJM, JAMA, Lancet, JACC, EHJ, etc.), (2) Write evidence-based editorials in Eric Topol's style from Ground Truth, (3) Create 500-word commentaries on clinical trials with PubMed citations, (4) Analyze trial importance using hybrid rules + LLM scoring, (5) Write editorials from full papers OR abstract-only scenarios, (6) Build thought leadership content for cardiologists, or (7) Synthesize recent cardiology advances for peers and referring physicians.

Prompt engineering guidance for Claude (Anthropic) model. Use when crafting prompts for Claude to leverage XML-style tags, long-context capabilities, extended thinking, and strong instruction following.

Store universal coding patterns into vector database. Auto-invokes after difficult tasks with broadly-applicable lessons. Trigger with "--store" or when user expresses frustration (strong learning signals). Uses true two-stage retrieval with MCP server v2.

This skill should be used when creating a Claude Code slash command. Use when users ask to "create a command", "make a slash command", "add a command", or want to document a workflow as a reusable command. Essential for creating optimized, agent-executable slash commands with proper structure and best practices.

Multi-model AI consultation and research using cursor-agent. Supports CONSULTATION (single query to specific model), RESEARCH (parallel multi-agent queries), and DELIBERATION (multi-round debate). Use when user says 'ask gpt', 'ask opus', 'consult gemini', 'do research on', 'deliberate on', or wants second opinions from AI models.

Use this skill when orchestrating multi-agent adversarial code reviews using Claude Code agent teams. Activates when conducting thorough code reviews, setting up review debate teams, spawning specialized review personas, or synthesizing findings from multiple reviewers. Provides spawn prompts, debate protocol, and synthesis templates.

Use this skill when orchestrating multi-agent adversarial planning debates before code is written. Activates when planning complex implementations, coordinating pre-implementation discovery, stress-testing implementation approaches, or producing validated plans for coding agents. Provides discovery prompts, debate protocol, and plan templates.

LangChain Deep Agents framework for building autonomous coding agents. Use for agent harness, backends, subagents, human-in-the-loop, long-term memory, middleware, and CLI-based agent development.

Comprehensive guide for authoring Claude Code extensions including skills (SKILL.md), hooks, slash commands, and subagents. Use when: creating skills, writing hooks, making slash commands, building subagents, learning Claude Code extensions.

Convert Markdown with LaTeX math into a native Google Doc whose equations are real, editable equation objects (not images, not literal $$ text). Use when: "/gdoc-math", "turn this markdown into a Google Doc with editable equations", "convert this LaTeX/formulas to a Google Doc", "make a gdoc from this .md with math", or converting an LLM/Gemini answer full of $...$ formulas into a proper editable Doc. NOT for: plain Markdown with no math (use /copy:gdocs), clipboard copy, Sheets/Slides, or editing an existing Doc's body.

Creates detailed, sectionized implementation plans through research, stakeholder interviews, and multi-LLM review. Use when planning features that need thorough pre-implementation analysis.

This skill should be used when the user asks to "set up gws", "install Google Workspace CLI", "connect Gmail to Claude", "manage Google Drive from terminal", "send email from CLI", "check my calendar", "create a Google Doc", "use gws", "Google Workspace agent", or mentions gws CLI, Google Workspace automation, or wants to interact with Gmail, Drive, Calendar, Sheets, Docs, Slides, Chat, Tasks, Meet, or Forms from the command line.

Turn a deadline, launch date, or delivery target into an executable Timeline Management Pack (deadline type + commitments, phase plan, milestone tracker, RAG cadence, scope/change control, stakeholder comms). Use for timeline/deadline/schedule/milestones.

Consult external LLMs (Gemini, OpenAI/Codex, Qwen) for second opinions, alternative plans, independent reviews, or delegated tasks. Use when a user asks for another model's perspective, wants to compare answers, or requests delegating a subtask to Gemini/Codex/Qwen.

Use when maintaining or optimizing OpenClaw workspace files — AGENTS.md, TOOLS.md, SOUL.md, USER.md, IDENTITY.md, HEARTBEAT.md, BOOT.md, MEMORY.md, and related checklists and memory files. Covers workspace auditing, token budget analysis, new agent workspace setup from scratch, memory distillation, and cross-file consistency reviews.

# prompt-manager

Optimizes QE agent performance through transfer learning, hyperparameter tuning, and pattern distillation across test domains. Use when improving agent accuracy, applying learned patterns to new projects, tuning quality thresholds, or implementing continuous improvement loops for AI-powered testing.

Create a self-contained review package of current work for external review by another AI model or human reviewer. Bundles relevant files with a contextual README and instructional prompt. Triggers: "review package", "create review package", "hand off for review", "get a second opinion", "external code review", "cross-model review", "package for review", "prepare code review". Accepts an optional focus area argument to scope the analysis.

Guidance for creating effective Claude Code plugin skills with proper structure, progressive disclosure, and content organization best practices.

Creates, audits, and validates agent skills following the Agent Skills open standard. Covers YAML frontmatter, description writing, progressive disclosure, directory structure, 9-phase audits, severity classification, and manifest-implementation drift detection. Use when creating new skills, writing skill descriptions, auditing existing skills, detecting version drift, validating implementations match manifests, or debugging skill discovery issues.

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

Intelligently triage bug reports and error messages by searching for duplicates in Jira and offering to create new issues or add comments to existing ones. When an agent needs to: (1) Triage a bug report or error message, (2) Check if an issue is a duplicate, (3) Find similar past issues, (4) Create a new bug ticket with proper context, or (5) Add information to an existing ticket. Searches Jira for similar issues, identifies duplicates, checks fix history, and helps create well-structured bug reports.

Use UMAP-learn for nonlinear dimensionality reduction, 2D/3D embeddings, clustering preprocessing, supervised or semi-supervised UMAP, DensMAP, AlignedUMAP, and Parametric UMAP workflows.

Persistent world-model harness that stress-tests ideas, strategies, and investments against 11 time horizons from 6 months to 50 years. Each horizon model is a deep (~10 page) analysis of geopolitics, technology, economics, society, environment, security, and wildcards stored at PAI_DIR/MEMORY/RESEARCH/WorldModels/. Three execution tiers: Fast (~2 min, single synthesizing agent), Standard (~10 min, 11 parallel horizon agents + RedTeam + FirstPrinciples), Deep (up to 1hr, adds per-horizon Research + Council). Four workflows: TestIdea (test any input across all 11 horizons, returns probability-weighted scenario matrix), UpdateModels (refresh model content with new research), ViewModels (read and summarize current state), TestScenario (test against alternative future models like great-correction-2027). Context files: ModelTemplate.md (structure for horizon model documents), OutputFormat.md (template for TestIdea results). Scenario models stored at WorldModels/Scenarios/. Orchestrates RedTeam, FirstPrinciples, Council, and Research internally. USE WHEN threat model, world model, test idea, test strategy, future analysis, test investment, time horizon analysis, update models, stress test against future, how does this hold up, long-term risk, what could go wrong over time, horizon analysis, crash scenario, view models, model status.

Add a redacted agent transcript section to GitHub PR or issue bodies during OpenClaw agent-created PR/issue workflows.

Reference for creating AI agent skills with SKILL.md structure. Use when building or improving modular skill files for Claude Code or Codex.

Configures AI agents (GitHub Copilot, Claude Code, Cursor, Codex) to understand a codebase and self-improve as the project evolves. Use when setting up a new project for AI-assisted development, onboarding AI agents to an existing repo, creating AGENTS.md, configuring MCP servers for code navigation, packaging project workflows as agent skills, or establishing self-improvement feedback loops. Covers AGENTS.md authoring, skill packaging, MCP configuration, context management, and living-documentation strategies.

Performs ARA Seal Level 2 semantic epistemic review on Agent-Native Research Artifacts, scoring six dimensions (evidence relevance, falsifiability, scope calibration, argument coherence, exploration integrity, methodological rigor) and producing a constructive, severity-ranked report with a Strong Accept-to-Reject recommendation. Use after Level 1 structural validation passes, when an ARA needs an objective epistemic critique before publication or release.

Performs ARA Seal Level 2 semantic epistemic review on Agent-Native Research Artifacts, scoring six dimensions (evidence relevance, falsifiability, scope calibration, argument coherence, exploration integrity, methodological rigor) and producing a constructive, severity-ranked report with a Strong Accept-to-Reject recommendation. Use after Level 1 structural validation passes, when an ARA needs an objective epistemic critique before publication or release.

Optimizes, improves, and debugs LLM prompts using production trace data, evaluations, and annotations. Extracts prompts from spans, gathers performance signal, and runs a data-driven optimization loop using the ax CLI. Use when the user mentions optimize prompt, improve prompt, make AI respond better, improve output quality, prompt engineering, prompt tuning, or system prompt improvement.

Claude Code extensibility: agents, skills, output styles. Capabilities: create/update/delete agents and skills, YAML frontmatter, system prompts, tool/model selection, resumable agents, CLI-defined agents. Actions: create, edit, delete, optimize, test extensions. Keywords: agent, skill, output-style, SKILL.md, subagent, Task tool, progressive disclosure. Use when: creating agents/skills, editing extensions, configuring tool access, choosing models, testing activation.

Iterative code refinement through plan → code → evaluate → refine cycles. Runs lint checks (ruff), tests (pytest), and structured self-evaluation each cycle, then diagnoses failures and refines. Decomposes complex tasks into sequential phases, iterates up to 3 times per phase (10 total). Use when: the main agent delegates a code task with 'MODE: MORE_EFFORT', the user selects 'More Effort' code generation mode, or the task explicitly requests iterative refinement for higher code quality. Do NOT use for single-pass code generation (Lite mode), experiment pipeline orchestration (use experiment-pipeline), or diagnosing a specific experiment failure (use experiment-craft).