Commands Reference

SpecFlow commands for PM orchestration, analysis, and development.

Primary Commands

These are the core commands you use most frequently:

Extended Commands

Additional commands for specialized workflows:

Command Details

/sf:pm - PM Orchestrator

The main entry point for all SpecFlow workflows. John (PM) orchestrates the entire feature lifecycle.

Usage:

/sf:pm "add logout button"         # Start new feature
/sf:pm --status                    # Check current state
/sf:pm --review                    # Review outputs

What it does:

1. Analyzes your request for pillar triggers
2. Creates feature folder in .specflow/features/
3. Routes to appropriate agents based on scope
4. Runs drift checkpoints after Dev and QA
5. Synthesizes requirements lock

/sf:analyst - Requirements Analysis

Mary analyzes scope and writes specifications with BOSS-compliant acceptance criteria.

Usage:

/sf:analyst    # Run within active feature

Modes:

• Mode 1: Scope assessment (creates 0-scope.md)
• Mode 1.5: Codebase analysis (creates 1.5-codebase-constraints.md)
• Mode 2: Specification (creates 1-spec.md)

When to use directly: To rerun scope assessment or update specifications.

/sf:architect - Architecture Design

Winston designs technical solutions that honor existing codebase patterns.

Usage:

/sf:architect    # Run within active feature

Output depth by scope:

When to use directly: To regenerate architecture after spec changes.

/sf:security - Security Analysis

Jordan applies STRIDE threat modeling to architecture designs.

Usage:

/sf:security    # Run within active feature

Triggers: Medium+ scope features involving:

• Authentication or authorization
• User data (PII, passwords, tokens)
• Payment processing
• Public API endpoints
• File uploads

Output: 3-security.md with STRIDE analysis and mitigations.

/sf:cost - Cost Analysis

Taylor estimates cloud resource costs using FinOps principles.

Usage:

/sf:cost    # Run within active feature

Output depth by scope:

Output: 4-cost.md with estimates and optimizations.

/sf:tea - Test Engineering Analysis

TEA creates Gherkin test scenarios and traceability matrices.

Usage:

/sf:tea    # Run within active feature

Output: 5-test-plan.md containing:

• Gherkin scenarios for all acceptance criteria
• Traceability matrix (AC to tests)
• Test count based on scope
• Flow recommendation (qa-first or dev-first)

/sf:dev - Development

Amelia implements features according to requirements lock.

Usage:

/sf:dev    # Run within active feature

Modes:

• Standard: Implement from requirements lock
• Fix Mode: Address review findings
• Drift Fix: Correct drift issues

Output: 6-dev-output.md with implementation summary.

/sf:qa - Quality Assurance

Quinn executes tests and verifies acceptance criteria.

Usage:

/sf:qa    # Run within active feature

Modes:

• Standard: Verify after Dev implements
• TDD: Write failing tests before Dev (qa-first)
• Fix Mode: Address review findings

Output: 7-qa-output.md with test results.

/sf:brainstorm - Structured Ideation

Uses 60+ brainstorming techniques for idea generation.

Usage:

/sf:brainstorm "how should we handle user onboarding?"

Standalone mode: Returns ideas document directly. PM-invoked mode: PM detects uncertainty, routes to brainstorm.

Output: 0.3-brainstorm.md when invoked via PM.

/sf:ux - UX Design Analysis

Sally analyzes user experience needs based on feature scope.

Usage:

/sf:ux "design the checkout flow"

Depth by scope:

Output: 1.6-ux-design.md with recommendations.

/sf:diagram - Architecture Visualization

Generates Excalidraw JSON for visual documentation.

Usage:

/sf:diagram --type flowchart "User login flow"
/sf:diagram --type wireframe "Dashboard layout"
/sf:diagram --type dataflow "API pipeline"
/sf:diagram --type architecture --from-arch

Output: Raw JSON for Excalidraw import.

/sf:docs - Technical Documentation

Paige generates documentation in various formats.

Usage:

/sf:docs --type readme "Document this API"
/sf:docs --type api --output docs/api.md
/sf:docs --type user-guide

Types: readme, api, user-guide, architecture, developer, release-notes

Output: 8-docs.md or specified output path.

/sf:review - Dynamic Review Orchestrator

Orchestrates code review with multiple review skills.

Usage:

/sf:review                    # Full review
/sf:review --security         # Security-focused review
/sf:review --lens=quality     # Specific lens

Available lenses: quality, security, performance, accessibility

Output: 8-review-output.md with findings.

/sf:dev-story - Develop Single Story

Implements a single story from the sprint.

Usage:

/sf:dev-story 1-1-auth-setup

Features:

• Checks story dependencies
• Updates sprint-status.yaml
• TDD with test evidence capture

Utility Commands

Planning Commands

Command Patterns

Starting a Feature

# The PM handles everything
/sf:pm "your feature description"

Checking Status

/sf:pm --status

Rerunning a Stage

/sf:analyst     # Redo scope/spec
/sf:architect   # Redo architecture
/sf:tea         # Redo test plan

Direct Agent Invocation

While PM routes automatically, you can invoke agents directly when you need to:

• Rerun a specific stage
• Debug workflow issues
• Learn agent capabilities

Tips

1. Start with PM: Always begin features with /sf:pm
2. Trust routing: Let PM handle agent sequencing
3. Check status: Use --status to see where you are
4. Review outputs: Read the artifacts agents produce
5. Scope first: Approve scope before deep analysis begins