agentdev-workflow-req-define

Orchestrates an interactive requirement-definition workflow producing structured draft-data requirement documents.

Updated Mar 7, 2026
One-click install
npx skills add https://github.com/yogata/agent-dev-flow --skill agentdev-workflow-req-define-yogata
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agentdev-workflow-req-define
Source: https://github.com/yogata/agent-dev-flow/tree/main/src/opencode/skills/agentdev-workflow-req-define
Command: npx skills add https://github.com/yogata/agent-dev-flow --skill agentdev-workflow-req-define-yogata

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Turning a vague feature idea or bug report into a well-structured, review-ready requirement document is slow and error-prone. This Skill controls the full req-define workflow of the AgentDevFlow plugin, guiding an AI agent through dialogue, existing-requirement matching, decision analysis, and draft generation so requirements are captured consistently and resumably. ## Core Features & Use Cases - 11-Step Workflow Control: Runs session-context detection, interactive requirement dialogue (HITL), existing REQ matching, requirement expansion, Decision judgment, draft-data generation, work_type/Scale classification, adversarial review, draft saving, and completion reporting. - Structured Draft Output: Produces .agentdev/drafts/req-draft-{topic-slug}.md in a structured draft-data YAML format containing operation_units, artifact_actions, test_strategy, and review_dispositions, consumed downstream by case-open, case-ready, and case-revise. - Resumable State Management: Persists agreed content into durable draft state so interrupted sessions resume from saved progress rather than conversation history. - Use Case: A developer describes a new feature in natural language; the Skill interviews them, checks existing REQ documents for APPEND/UPDATE candidates, derives boundary risks and a test strategy, runs an adversarial review, and saves a requirement draft ready for case-open. ## Quick Start Invoke the /agentdev/req-define command with a description of your feature or bug fix and let the workflow guide you through requirement definition.

Frequently Asked Questions about agentdev-workflow-req-define

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

FAQPage Schema
How do I define a software requirement with an AI agent workflow?▼

Run the /agentdev/req-define command with a natural-language description of your feature or bug fix. The workflow interviews you, matches existing requirements, derives risks and a test strategy, and saves a structured draft under .agentdev/drafts/.

What output format does the req-define workflow produce?▼

It produces a req-draft-{topic-slug}.md file containing a structured draft-data fenced YAML block with operation_units, artifact_actions, test_strategy, and review_dispositions. Downstream commands case-open, case-ready, and case-revise consume this draft.

Can I resume an interrupted requirement definition session?▼

Yes. Agreed content is persisted incrementally into the draft-data draft, so resumption reconstructs state from durable files rather than conversation history. Unresolved questions and stop reasons are recorded in the draft fields.

Does req-define write implementation code or run git commands?▼

No. It operates only in the dialogue phase and may create or edit files solely under .agentdev/drafts/. Git commands are forbidden, and referenced input files such as requirement units are read-only.

When is the adversarial review step skipped in req-define?▼

The review is default-on but skipped when Scale is L0 with no Decision targets and no semantic decisions. A user explicit request forces it to run regardless, and unresolved user judgment items block progression to draft saving.