LLM 與 Agent

First-draft writing assistant for Crucible-planned novels. Use when author has completed planning (crucible-planner) and outlining (crucible-outliner) and wants to write the actual prose. Handles scene-by-scene drafting, style matching, continuity tracking, and hallucination prevention. Triggers on "write my novel," "draft chapter X," "start writing from my outline," "help me write my book," or when user has Crucible outline and wants prose output. Works with any Crucible-planned story.

IMMEDIATELY load this skill if you read ANY file containing "taskagent" (PLAN.md, org/projects, docs, etc). Also load for multi-session work or when user mentions agent tasks. For simple single-session tasks without taskagent references, use TodoWrite instead.

Analyzes Rails applications and generates comprehensive upgrade reports with breaking changes, deprecations, and step-by-step migration guides for Rails 7.0 through 8.1.1. Use when upgrading Rails applications, planning multi-hop upgrades, or querying version-specific changes.

Expert on Aptos Framework (0x1 standard library) - account, coin, fungible_asset, object, timestamp, table, event, vector, string, option, error, and other core modules. Triggers on keywords aptos framework, 0x1, account module, table, smarttable, event, timestamp, randomness, aggregator, resource account

Research-backed creation and enhancement of opencode agents. Use when creating new agents from requirements, improving underperforming agents, or designing agent configurations. Triggers on "create an agent for X", "improve my Y agent", "design a specialized agent", "my agent isn't working well".

Expert on Move programming language - abilities (copy/drop/store/key), generics, phantom types, references, global storage operations, signer pattern, visibility modifiers, friend functions, inline optimization, and advanced type system. Triggers on keywords move language, abilities, generics, phantom type, borrow global, signer, friend, inline, type parameter

Expert context manager specializing in information storage, retrieval, and synchronization across multi-agent systems. Masters state management, version control, and data lifecycle with focus on ensuring consistency, accessibility, and performance at scale.

Create slash commands for Claude Code with $ARGUMENTS handling, agent invocation patterns, and template best practices. Reference for building user-triggered workflow shortcuts.

Expert on Claude Code CLI, skills, commands, hooks, plugins, MCP, settings, and workflows. Triggers on claude code, cli, skill, command, hook, plugin, mcp, slash command, settings

Configure OpenCode settings and rules. Use when user asks to "configure opencode", "set up opencode", "create opencode config", "edit opencode.json", "set up AGENTS.md", "configure providers", "set up permissions", or needs help with OpenCode configuration files.

Expert test automation engineer specializing in building robust test frameworks, CI/CD integration, and comprehensive test coverage. Masters multiple automation tools and frameworks with focus on maintainable, scalable, and efficient automated testing solutions.

Operational AI search engine optimization (AEO/GEO/LLMO) for ranking in ChatGPT, Perplexity, Claude, Gemini, and other LLM-powered search platforms with technical setup, content strategies, and measurement frameworks.

Expert on integrating Shelby Protocol decentralized storage into applications. Helps with use case evaluation, architecture design, migration strategies, cost optimization, performance tuning for video streaming, AI training, data analytics, content delivery, and read-heavy workloads. Triggers on keywords integrate Shelby, decentralized storage integration, video streaming storage, AI training data, data analytics storage, migration to Shelby, storage architecture, content delivery, Shelby use case.

Systematic research methodology for building comprehensive, current knowledge on any topic. Requires web_search tool. Use when questions require thorough investigation, recent developments post-cutoff, synthesis across multiple sources, or when Claude's knowledge may be outdated or incomplete. Triggered by "Research", "Investigate", "What's current on", "Latest info on", complex queries needing validation, or technical topics with recent changes.

Expert on Anthropic Claude API, models, prompt engineering, function calling, vision, and best practices. Triggers on anthropic, claude, api, prompt, function calling, vision, messages api, embeddings

Create OpenCode plugins using @opencode-ai/plugin SDK. Use when user wants to BUILD a plugin in .opencode/plugin/ or ~/.config/opencode/plugin/ - for custom tools, event hooks, LLM parameter control, auth providers, or tool execution interception. DO NOT use when modifying OpenCode source code (packages/plugin, packages/opencode, etc.) or working on OC internals - those are core development, not plugin creation.

Create and configure Claude Code agents with YAML frontmatter, tool selection, model specification, and naming conventions. Reference for building specialized AI subagents that handle complex, multi-step tasks.

Expert knowledge synthesizer specializing in extracting insights from multi-agent interactions, identifying patterns, and building collective intelligence. Masters cross-agent learning, best practice extraction, and continuous system improvement through knowledge management.

Expert LLM architect specializing in large language model architecture, deployment, and optimization. Masters LLM system design, fine-tuning strategies, and production serving with focus on building scalable, efficient, and safe LLM applications.

Expert on Move Prover formal verification - specification language (MSL), preconditions, postconditions, invariants, aborts_if, ensures, requires, modifies, emits, global invariants, schema patterns, quantifiers, helper functions, pragma directives, verification strategies, and debugging proofs. Triggers on keywords move prover, formal verification, spec, invariant, ensures, requires, aborts_if, precondition, postcondition, quantifier, schema, pragma