Skip to content
Agent Orchestration
Skill

/design-doc-generate

Generate design documents for Validate mode work items (Checkpoint 1). Required for high-complexity items.

BOOST
From plugin
specsmd
22115 skills19 agents22 commands
Install
$ npx -y skills add fabriqaai/specs.md --skill design-doc-generate --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/design-doc-generate

Context preview

The summary Claude sees to decide when to auto-load this skill.

Generate design documents for Validate mode work items (Checkpoint 1). Required for high-complexity items.

SKILL.md

design-doc-generate.SKILL.md
name: design-doc-generate
description: Generate design documents for Validate mode work items (Checkpoint 1). Required for high-complexity items.
version: 1.0.0

<objective> Generate design documents for Validate mode work items (Checkpoint 1). </objective>

<triggers>

  • Work item has complexity: high
  • Work item has mode: validate
  • Design review required before implementation

</triggers>

<degrees_of_freedom> **LOW** — Follow the design doc structure precisely. Decisions must have rationale. </degrees_of_freedom>

<llm critical="true"> <mandate>Design doc MUST be approved before implementation</mandate> <mandate>Document DECISIONS with RATIONALE, not just choices</mandate> <mandate>Keep concise — enough detail to implement, no more</mandate> <mandate>Include risks upfront — don't hide complexity</mandate> </llm>

<flow> <step n="1" title="Analyze Work Item"> <action>Read work item from .specs-fire/intents/{intent-id}/work-items/{id}.md</action> <action>Identify key design decisions needed</action> <action>Assess domain modeling needs</action> <action>Identify integration points</action> </step>

<step n="2" title="Gather Context"> <action>Review project standards (.specs-fire/standards/)</action> <action>Check existing codebase patterns</action> <action>Identify similar implementations to reference</action> </step>

<step n="3" title="Draft Key Decisions"> <action>For each decision point:</action> <substep>Identify options considered</substep> <substep>Evaluate trade-offs</substep> <substep>Select recommended choice</substep> <substep>Document rationale</substep>

<output_format> | Decision | Choice | Rationale | |----------|--------|-----------| | ... | ... | ... | </output_format> </step>

<step n="4" title="Define Domain Model" if="has_domain_complexity"> <action>Identify entities (things with identity)</action> <action>Identify value objects (immutable values)</action> <action>Identify domain events (if event-driven)</action> <action>Map relationships</action> </step>

<step n="5" title="Design Technical Approach"> <action>Create component diagram (ASCII)</action> <action>Define API contracts (if applicable)</action> <action>Specify database changes (if applicable)</action> <action>Document data flow</action> </step>

<step n="6" title="Identify Risks"> <action>List potential risks</action> <action>Assess impact (high/medium/low)</action> <action>Propose mitigations</action>

<output_format> | Risk | Impact | Mitigation | |------|--------|------------| | ... | ... | ... | </output_format> </step>

<step n="7" title="Create Implementation Checklist"> <action>Break down into implementation steps</action> <action>Order by dependency</action> <action>Keep granular but not excessive</action> </step>

<step n="8" title="Present Design Doc"> <checkpoint message="Design document ready for review"> <output>

Design: {work-item-title}

Summary

{brief description}

Key Decisions

{decisions table}

Technical Approach

{component diagram, API contracts}

Risks

{risks table}

Implementation Checklist

{ordered steps}

--- This is Checkpoint 1 of Validate mode.

Approve design? [Y/n/edit] </output> </checkpoint> </step>

<step n="9" title="Handle Response"> <check if="response == y"> <action>Generate design doc using template: templates/design.md.hbs</action> <action>Save to: .specs-fire/intents/{intent-id}/work-items/{id}-design.md</action> <action>Mark checkpoint 1 as passed</action> <output> Design approved. Ready for implementation planning.

Route to Builder for Checkpoint 2 (implementation plan)? [Y/n] </output> </check> <check if="response == edit"> <ask>What changes are needed?</ask> <action>Incorporate feedback</action> <goto step="8"/> </check> <check if="response == n"> <output>Design rejected. What concerns need to be addressed?</output> <action>Gather feedback, revise approach</action> <goto step="3"/> </check> </step> </flow>

<output_artifacts>

| Artifact | Location | Template | |----------|----------|----------| | Design Doc | `.specs-fire/intents/{intent-id}/work-items/{id}-design.md` | `./templates/design.md.hbs` | </output_artifacts>

<success_criteria> <criterion>Work item analyzed for design decisions</criterion> <criterion>Key decisions documented with rationale</criterion> <criterion>Domain model defined (if applicable)</criterion> <criterion>Technical approach specified</criterion> <criterion>Risks identified with mitigations</criterion> <criterion>Implementation checklist created</criterion> <criterion>Design doc approved at checkpoint</criterion> <criterion>Design doc saved to correct location</criterion> </success_criteria>

Read more
Ships withspecsmd

AI-native development framework with pluggable flows for every use case. Choose the flow that matches your project needs: Ideation for creative brainstorming, Simple for quick specs, FIRE for adaptive execution, or AI-DLC for full methodology with DDD.

Get the whole plugin

Other skills on specsmd.