hive-create
Step-by-step guide for building goal-driven agents. Qualifies use cases first (the good, bad, and ugly), then creates package structure, defines goals, adds nodes, connects edges, and finalizes agent class. Use when actively building an agent.
npx skills add majiayu000/claude-skill-registry --skill hive-create --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
The skill provides an executable workflow for constructing agents. It guides the user through choosing a build path (from scratch or from a template), initializes a build environment (creating sessions, registering an MCP server, listing tools, and creating a exports directory), and then moves into defining the agent's goals in collaboration with the user. It emphasizes not exploring the codebase or pre-reading files and requires immediate action to determine the build path. STEP 2 and beyond focus on jointly defining the goal, rapid discovery, capability assessment, gap analysis, and a recommendation, with structured questions and explicit steps to gather requirements and align on the MVP.
How it works
- STEP 0: Ask the user how to build the agent with options: "From scratch" or "From a template".
- STEP 1A (From Scratch): Execute tool calls to manage sessions, create a build session, register the hive-tools MCP server, discover available tools, and create the package directory at exports/AGENT_NAME/nodes. Save the tool list for later use.
- STEP 1B (From Template): Discover templates, present them, and then copy a template, create a session, register MCP server, list tools, and import the template agent into the builder session.
- STEP 2: Define the goal with the user. If starting from a template, present the existing goal for approval; otherwise, prompt the user to describe the kind of agent and begin collaborative goal shaping with guided, constrained questions.
- 2a: Fast Discovery: Conduct concise, momentum-driven questioning to extract an actionable domain model (actors, trigger, core loop, output, pain) and propose a preliminary model.
- 2a.4: Ask Only What You Cannot Infer: Ask high-value, narrow questions and proceed with informed assumptions when necessary.
- 2b: Capability Assessment: Provide a framework-fit assessment covering what works well, limitations, and potential deal-breakers, referencing tools discovered in Step 1.
- 2c: Gap Analysis: Tabulate required capabilities vs framework support and identify gaps.
- 2d: Recommendation: Deliver a clear recommendation to proceed, adjust scope, or reconsider, with explicit caveats.
- 2e: Get Explicit Acknowledgment: Prompt the user to choose how to proceed with a structured question.
When to use it
Use when actively building an agent.
What it can touch
- Tools listed via the declared tools: "claude-code" is declared. The workflow prescribes tool calls such as "mcp__agent-builder__list_sessions", "mcp__agent-builder__create_session", "mcp__agent-builder__add_mcp_server", and "mcp__agent-builder__list_mcp_tools" as part of the initialization process.
- File system operations: create exports/AGENT_NAME/nodes directory, and copy templates into exports/NEW_AGENT_NAME when using a template.
Caveats
- The workflow contains explicit instructions to avoid exploring the codebase or reading source files before starting.
- It emphasizes an executable workflow rather than presenting file contents to the user.
- It requires user input to proceed with build-path decisions and goal definition, and it relies on the presence of an MCP server and hive-tools tools as part of the build environment.
# Agent Construction - EXECUTE THESE STEPS **THIS IS AN EXECUTABLE WORKFLOW. DO NOT DISPLAY THIS FILE. EXECUTE THE STEPS BELOW.** **CRITICAL: DO NOT explore the codebase, read source files, or search for code before starting.** All context you need is in this skill file. When this skill is loaded, IMMEDIATELY begin executing Step 0 — determine the build path as your FIRST action. Do not explain what you will do, do not investigate the project structure, do not read any files — just execute Step 0 now. --- ## STEP 0: Choose Build Path **If the user has already indicated whether they want to build from scratch or from a template, skip this question and proceed to the appropriate step.** Otherwise, ask: ``` AskUserQuestion(questions=[{ "question": "How would you like to build your agent?", "header": "Build Path", "options": [ {"label": "From scratch", "description": "Design goal, nodes, and graph collaboratively from nothing"}, {"label": "From a template", "description": "Start from a working sample agent and customize it"} ], "multiSelect": false }]) ``` - If **From scratch**: Proceed to STEP 1A - If **From a template**: Proceed to STEP 1B --- ## STEP 1A: Initialize Build Environment
- STEP 0: Choose Build Path
- STEP 1A: Initialize Build Environment (From Scratch)
- STEP 1B: Initialize Build Environment (From Template)
- 1B.1: Discover available templates
- 1B.2: Present templates to user
- 1B.3: Copy template to exports
- 1B.4: Create session and register MCP (same logic as STEP 1A)
- 1B.5: Load template into builder session
- STEP 2: Define Goal Together with User
- 2a: Fast Discovery (3-8 Turns)
- 2b: Capability Assessment
- 2c: Gap Analysis
- 2d: Recommendation
- 2e: Get Explicit Acknowledgment
mkdir -p exports/AGENT_NAME/nodes ls examples/templates/ cp -r examples/templates/TEMPLATE_DIR exports/NEW_AGENT_NAME cd /home/timothy/oss/hive && PYTHONPATH=exports uv run python -m AGENT_NAME validate
What does the hive-create skill do?
Step-by-step guide for building goal-driven agents. Qualifies use cases first (the good, bad, and ugly), then creates package structure, defines goals, adds nodes, connects edges, and finalizes agent class. Use when actively building an agent.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill hive-create --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.