xoai/design

• PERSONA: Read sage/core/agents/analyst.persona.md for your mindset.

Design Workflow

Shape solutions. Reads research context, produces specs ready for build.

Auto-Pickup

Scan .sage/work/ for design initiatives with status: in-progress. This scan is MANDATORY — check the DISK.

Scan .sage/docs/ for upstream research artifacts: jtbd-.md, ux-audit-.md, opportunity-.md, user-interview-.md

If research found: "Sage: Found research context — [list]. Using as design input."

If design artifacts exist: resume from current phase.

Read .sage/decisions.md for context. Read handoff field in the most recent artifact if present.

Step 1: Scope the Design (Zone 1)

If research context was found:

Sage → design workflow. Found research: [list of .sage/docs/ artifacts].

[1] Feature design — UX brief → detailed spec (2 steps) [2] Content/copy — UX writing (1 step) [3] Product requirements — PRD (1 step) [4] Comprehensive — brief → spec → copy (3 steps)

Pick 1-4, type / for commands, or describe what you need.

If no research context:

Sage → design workflow. No prior research found.

[1] Feature design — UX brief → detailed spec (2 steps) [2] Content/copy — UX writing (1 step) [3] Product requirements — PRD (1 step) [4] Research first — type /research to understand before designing

Pick 1-4, type / for commands, or describe what you need.

Step 2: Execute Skill Chain

Based on scope, load and execute skills in sequence:

| Scope | Skill Chain | |-------|-------------| | Feature | ux-brief → ux-specify | | Content | ux-writing | | PRD | prd | | Comprehensive | ux-brief → ux-specify → ux-writing |

For each skill in the chain:

1. Announce: "Sage: Starting [skill name]."
2. Load sage/skills/[skill]/SKILL.md and follow its process
3. If research context exists, reference it during execution
4. Save output to .sage/work/YYYYMMDD-slug/[artifact].md with standard frontmatter (title, status, phase)
5. Present per-artifact checkpoint (Zone 2):

Sage: [Artifact] saved to .sage/work/YYYYMMDD-slug/[name].md Decision: [key design decision]. (appended to decisions.md)

[A] Approve — continue to next step [R] Revise

Pick A/R, or tell me what to change.

6. On approval, proceed to next skill in chain

Step 3: Design Completion (Zone 2)

🔒 DESIGN CHECKPOINT:

Self-check (observable conditions):

• [ ] Design artifacts exist in .sage/work/
• [ ] If research context was available, findings are reflected
• [ ] Key design decisions are documented

Sage: Design complete.

Artifacts: .sage/work/YYYYMMDD-slug/brief.md .sage/work/YYYYMMDD-slug/spec.md

Decision: [key design decisions]. (appended to decisions.md)

[A] Approve [R] Revise [V] → /review [N] New session → /build

Pick A/R/V/N, or tell me what to change.

On approval: update artifact frontmatter to status: completed. Write handoff field:

handoff: |
  Key decisions: [design choices and rationale]
  Research context: [what research informed this design]
  Open questions: [what needs resolution during build]
  Next agent should: [guidance for implementation]

Append to decisions.md (Rule 7).

Step 4: Next Step (Zone 3)

Next steps: /build — spec → plan → implement → verify (reads your design) /review — independent evaluation of the design /reflect — review design decisions before building

Type a command, or describe what you want to do next.

Quality Criteria

Good design output:

• Grounded in research when available (references findings)
• Decisions have explicit rationale (not just "I chose X")
• User flows are concrete (screens, states, transitions)
• Edge cases are identified and handled
• The spec is implementable — a developer can build from it

Rules

• Load research context if it exists in .sage/docs/
• Save all artifacts to .sage/work/ with frontmatter
• Checkpoints between each skill in the chain
• Handoff field on completion for the next agent
• If no research exists and scope is complex, suggest /research first