Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Browse local OpenCode history: sessions, messages, plans, prompt history, and prior decisions. Use when the user says history, previous session, last time, remember, recall, plans, prior work, or when resuming/debugging repeated work where earlier OpenCode context may help. Do not use for fresh tasks or when current files/git already answer the question.
A fast single-pass scoping map of a MEL/SRHR question: which perspectives it demands, where those perspectives conflict, and whether it is worth a full orchestration, ending in a triage verdict, never citations or findings. Use when Ane wants a question, proposal angle, learning question, or evaluation framing scoped or triaged before committing tokens to a full run. Not an interview of Ane (/grill-mel), not the deliverable or any sourced evidence (/ann, /researcher, /evidence-synthesis), and not a multi-agent debate (/decision-debate).
Advisor mode. Consult a stronger (or different) model at key checkpoints: before major decisions, when stuck on an error, and before declaring a task done. Use when the user types /advisor, asks to turn the advisor on or off, picks an advisor model, or explicitly asks for a second opinion from a stronger or different model. Never enable it on your own.
Orchestrate agile development workflows for Gitea repositories using the tea CLI. Use when working with Gitea-hosted repos and asking to 'run the workflow', 'continue working', 'what's next', 'complete the task cycle', 'start my day', 'end the sprint', 'implement the next task', or wanting guided step-by-step development assistance. Keywords: workflow, orchestrate, agile, task cycle, sprint, daily, implement, review, PR, standup, retrospective, gitea, tea.
Read and dig into Scrawl screen captures, where someone recorded their screen while talking, clicking and drawing red marks on it. Use whenever a message contains a Scrawl capture (a line like `[scrawl 0:12 · App “Window” · /path/to/….scrawl]`, a `.scrawl` bundle path, or a contact-sheet.png from one), when the user says "look at my scrawl" or refers to something they just recorded or drew, or when you need the `scrawl` CLI to zoom into a moment, pull frames, send a capture to a session, or export one as video.
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
Show the user real screenshots of the work as it progresses, without being asked. Capture the start, key moment, and result of every visual change; send them as files with captions as soon as each is ready, and look at every image before sending it. Use on any task with a visible result (UI, pages, games, 3D, charts, motion), when the user says “show me,” “screenshots,” “keep giving me screenshots,” or “where are the screenshots?”, and before any report that claims a visual change works.
When the user wants to build or improve a sales bot's ability to learn from human corrections and feedback. Also use when the user mentions "human feedback," "HITL," "human-in-the-loop," "feedback learning," or "correction learning."
Leverage OpenAI Codex/GPT models for autonomous code implementation, code review, and plan review. Triggers: "codex", "use gpt", "gpt-5", "let openai", "full-auto", "adversarial review", "second opinion review", "用codex", "让gpt实现", "对抗式审查", "让codex审查计划", "第二意见". Use this skill whenever the user wants to delegate coding tasks to OpenAI models, run code or plan reviews via codex, get a second-opinion review from a different model, or execute tasks in a sandboxed environment.
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
When the user wants to build or improve a sales bot's ability to reply fast enough to feel real-time but not unnaturally instant. Also use when the user mentions "response speed," "reply timing," "bot response delay," "natural pacing," or "message timing."
When the user wants to build or improve a sales bot's ability to dynamically choose conversation paths based on prospect responses rather than following rigid scripts. Also use when the user mentions "dynamic conversations," "branching logic," "conversation flows," "adaptive scripts," or "conditional responses."
Use when the user wants to build or improve a sales bot's ability to learn from deal outcomes. Also use when the user mentions “feedback loop,” “outcome learning,” “win/loss learning,” “continuous improvement,” or “model training.”
When the user wants to build or improve a sales bot's ability to know when it's uncertain. Also use when the user mentions "confidence scoring," "response confidence," "bot uncertainty," "escalation triggers," or "hedging responses."
Adopt repository-level harness engineering for coding agents. Use when a user wants to prevent repeated AI coding-agent mistakes by turning failures into durable instructions, drift checks, regression tests, failure memory, and adoption reports tailored to the target repository.
When the user wants to build or improve a sales bot's ability to anticipate likely responses. Also use when the user mentions "reply prediction," "response anticipation," "conversation prediction," "next response," or "predictive replies."
Use this skill when the user wants to turn a recurring browser workflow into a reusable, parameterized agent skill, especially when the task has concrete inputs and a clear output such as scheduled scrapes, form submissions, data extraction, monitoring flows, price probes, or login-gated reports. Do not use for one-off web tasks; use steel-browser.
Use when the user wants to build or improve a sales bot's ability to handle unexpected inputs without breaking the conversation. Also use when the user mentions "handling errors," "unexpected input," "bot failures," "graceful degradation," "fallback responses," or "error recovery."
Build MCP servers in Python with FastMCP to expose tools, resources, and prompts to LLMs. Supports storage backends, middleware, OAuth Proxy, OpenAPI integration, and FastMCP Cloud deployment. Prevents 30+ errors. Use when: creating MCP servers, or troubleshooting module-level server, storage, lifespan, middleware, OAuth, background tasks, or FastAPI mount errors.
Use when the user wants to build or improve a sales bot's ability to maintain coherent conversations across dozens of exchanges without losing context. Also use when the user mentions "conversation history," "remembering context," "long conversations," "maintaining thread," or "context window management."
Use when the user wants to build or improve a sales bot's ability to gracefully redirect when prospects ask about unrelated things. Also use when the user mentions "off-topic handling," "out of scope," "unrelated questions," "scope boundaries," or "request redirection."
When the user wants to build or improve a sales bot's ability to clarify vague or ambiguous responses before proceeding. Also use when the user mentions "clarifying questions," "disambiguation," "unclear responses," "vague answers," or "asking for clarity."
When the user wants to build or improve a sales bot's ability to seamlessly connect prospects to live reps with full context passed along. Also use when the user mentions "warm handoff," "transfer to human," "live rep handoff," "context handover," or "seamless transfer."
Anti-detection browser automation using Camoufox (Firefox fork with C++ fingerprint spoofing). Use when standard browser tools are blocked by Cloudflare, Akamai, or bot detection. Triggers include “stealth browse”, “anti-detection”, “bypass bot”, “camofox”, “blocked by Cloudflare”, scraping protected sites (X/Twitter, Amazon, Product Hunt), or when agent-browser/playwright fails with bot detection errors.
Package entire code repositories into single AI-friendly files using Repomix. Capabilities include pack codebases with customizable include/exclude patterns, generate multiple output formats (XML, Markdown, plain text), preserve file structure and context, optimize for AI consumption with token counting, filter by file types and directories, add custom headers and summaries. Use when packaging codebases for AI analysis, creating repository snapshots for LLM context, analyzing third-party libraries, preparing for security audits, generating documentation context, or evaluating unfamiliar codebases.
Understand and customize the local Trellis architecture inside a user project. Use when modifying .trellis plus platform hooks, settings, agents, skills, commands, prompts, workflows, the channel runtime (trellis channel), bundled runtime agents under .trellis/agents/, selectable workflow templates, registry-backed spec refresh, cross-session memory (trellis mem) generated by trellis init, or AI-facing bundled skills (trellis-channel, trellis-session-insight, trellis-spec-bootstrap) and bundled-skill auto-dispatch flow.
Claude Code Bash tool patterns with hooks, automation, git workflows. Use for PreToolUse hooks, command chaining, CLI orchestration, custom commands, or encountering bash permissions, command failures, security guards, hook configurations.
Use this skill when the user asks to “add MCP server”, “integrate MCP”, “configure MCP in plugin”, “use .mcp.json”, “set up Model Context Protocol”, “connect external service”, mentions “${CLAUDE_PLUGIN_ROOT} with MCP”, or discusses MCP server types (SSE, stdio, HTTP, WebSocket). Provides comprehensive guidance for integrating Model Context Protocol servers into Claude Code plugins for external tool and service integration.
Guide the user through a quadrant walk that maps the unknowns of a task — open by listing the known knowns, then work through known unknowns, unknown knowns, and unknown unknowns one stage at a time, ending with a complete four-quadrant map in the user's hands. Use when a request is ambiguous or underspecified, the codebase or domain is unfamiliar, the user will "know it when they see it", a reference implementation must be understood before porting, mid-build deviations from the plan need capturing, or a finished change needs buy-in or verified understanding before merge. Pairs with [write-spec](../write-spec/SKILL.md) — walk the quadrants to burn off fog before slicing, and feed the finished map into the spec.
Expert guidance on the Google Agent Development Kit (ADK) for Python. Use this skill when the user asks about building agents, using tools, streaming, callbacks, tutorials, deployment, or advanced architecture with the Google ADK in Python.
This skill should be used for persistent semantic memory in agent systems: cross-session knowledge retention, entity tracking, temporal validity, graph or vector retrieval, memory consolidation, and memory benchmark selection. Route file-backed scratchpads to filesystem-context, handoff summaries to context-compression, and token-efficiency tactics to context-optimization
Use this skill when the user asks to “apply skill improvements”, “update skill from plan”, “execute improvement plan”, “fix skill issues”, “implement skill recommendations”, or mentions applying improvements from quality review reports. Reads improvement-plan-{name}.md files generated by skill-quality-reviewer and intelligently merges and executes the suggested changes to improve Claude Skills quality.
Every product will be AI-powered. The question is whether you'll build it right or ship a demo that falls apart in production. This skill covers LLM integration patterns, RAG architecture, prompt engineering that scales, AI UX that users trust, and cost optimization that doesn't bankrupt you. Use when: keywords, file_patterns, code_patterns.
Universal context management and planning system. PROACTIVELY activate for: (1) ANY complex task requiring planning, (2) Multi-file projects/websites/apps, (3) Architecture decisions, (4) Research tasks, (5) Refactoring, (6) Long coding sessions, (7) Tasks with 3+ sequential steps. Provides: optimal file creation order, context-efficient workflows, extended-thinking delegation (~23x context efficiency), passive deep-analysis architecture, progressive task decomposition. Environment-agnostic — works for Web, API, and Claude Code CLI.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.