expert-team

Generates persistent two-layer documentation packages for business modules using parallel expert sub-agents.

2|Updated Jun 8, 2026
One-click install
npx skills add https://github.com/HACK-WU/skills --skill expert-team-hack-wu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: expert-team
Source: https://github.com/HACK-WU/skills/tree/main/skills/expert-team
Command: npx skills add https://github.com/HACK-WU/skills --skill expert-team-hack-wu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Repeatedly re-reading the same business module's source code wastes effort during design, refactoring, debugging, and onboarding. This Skill deeply analyzes a single business module once and persists the result as a reusable "module expert" asset package, so future work can consult the documentation instead of re-tracing code. ## Core Features & Use Cases - Two-layer documentation: Produces a black-box contract layer (capability contracts, usage flows, real code examples, data flow, key decisions) plus a white-box implementation layer (architecture, implementation, data flow, models, API, tests, ops) under .module-experts/. - Scalable organization: Supports sub-experts for large modules, topics for very large modules (topic → expert → sub-expert), and batched creation for modules too big to process at once. - Quality pipeline: Runs plan preview with user confirmation, parallel sub-agent research, format self-checks (R1–R7 / CR1–CR10), auto-review, index updates, and knowledge-base memory writes. - Use Case: Before refactoring a payment module, ask the AI to deeply analyze it; the Skill produces a payment expert package with API contracts, runnable usage examples, architecture maps, and test status that any future session can reuse. ## Quick Start Ask the AI to deeply analyze the payment module and build a module expert package for it.

Frequently Asked Questions about expert-team

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create deep documentation for a single business module?

Invoke the expert-team workflow with a clear module root path. It scans the module, presents a creation plan for confirmation, dispatches parallel sub-agents per aspect, and writes contract and implementation documents into .module-experts/.

What is the difference between the contract layer and implementation layer?

The contract layer (C0–C5 files in the expert root) is a black-box usage guide covering capabilities, parameters, and real code examples. The implementation layer (implementation/01–07) is a white-box map of architecture, data flow, and tests used to navigate the source code.

When should a module be split into sub-experts or a topic?

Split into sub-experts when the module has clear functional subdomains and one expert's contract would be too large. Create a topic only for very large modules (e.g., over 500 source files) needing a three-level topic → expert → sub-expert structure; sub-experts cannot nest further.

Which modules should not get an expert package?

Skip shared or cross-cutting code such as utility libraries, style conventions, and infrastructure, which belong in project memory instead. Also skip small or simple modules where reading the code directly or a lightweight walkthrough is sufficient.

How does the skill keep expert documentation from going stale?

Each expert records a baseline git commit in agent.md and INDEX.md. A companion lookup skill compares that baseline against HEAD to detect changed files and performs limited incremental updates without refreshing the baseline.

What happens if the knowledge base (ki) is unavailable during creation?

The memory-writing steps are skipped and the omission is noted in the final acceptance report. Expert package creation and disk output are not blocked by the missing knowledge base.