Install
$ agentstack add skill-akoliteza-hermes-agent-idea-workflow-idea-to-ui-design-brief ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Idea to UI Design Brief
Overview
Use this skill as an optional Full-mode stage inside the idea workflow. It converts a product/design direction into a practical UI design brief that can guide image-generation concepts, HTML prototypes, or later Superpowers implementation work.
This skill should not run in Lite mode unless the user explicitly asks for UI design. Lite mode should keep the previous simple idea-capture behavior.
The UI brief is a bridge between product thinking and implementation. It should define what the interface should look, feel, and behave like without forcing premature code architecture.
When to Use
Use when the user asks to:
- explore possible UI designs for an idea;
- add a dedicated UI/design stage to the idea workflow;
- generate image prompts for app screenshots, dashboards, landing pages, or product surfaces;
- create a UI brief before giving the idea to a coding agent;
- critique or redesign an already-built app after screenshots exist;
- turn a design doc into screen-by-screen UI requirements.
Do not use when:
- the user only wants a quick Lite idea note;
- the idea is not yet clear enough to identify target users and core behavior;
- the user wants production frontend code immediately — use implementation skills after the UI direction is chosen;
- the user wants a formal token spec only — use
design-mdif available.
Artifact Placement
For Full mode idea packages, add this artifact after the design doc and before implementation planning:
ideas//
README.md
00-idea-capture.md
01-design-doc.md
02-ui-design-brief.md
03-implementation-spec.md
04-agent-build-handoff.md
05-spec-review.md
Fallback rule: if the user skips this stage, preserve the old workflow numbering and behavior, or mark UI decisions as assumptions in the implementation spec. Do not block the workflow merely because no UI brief exists.
Required Inputs
Before writing the UI brief, collect or infer:
- product purpose and target user;
- core user flow;
- main screens or surfaces;
- platform target: web, desktop, mobile web, native mobile, or multi-surface;
- desired feel: calm, command-center, playful, professional, dense, minimal, etc.;
- visual constraints: brand, existing repo/design system, accessibility needs, dark/light mode;
- important states: empty, loading, error, success, first-run, repeat-use.
If these are missing and the user has not forced progression, ask one concise question at a time.
UI Brief Structure
Create 02-ui-design-brief.md with:
# — UI Design Brief
## Purpose
## Product Feel
## Design Principles
## Primary Surfaces / Screens
## Screen-by-Screen Notes
## Layout / Information Architecture
## Component Inventory
## Key User Flows
## States to Design
## Visual Direction
## Accessibility and Responsiveness
## Content / Copy Notes
## Optional Image-Generation Concepts
## Selected Direction
## Open Questions
## Handoff Notes for Implementation
Optional Image-Generation Pass
Image generation is optional and should be treated as concept exploration, not the implementation source of truth.
Use image generation when:
- the user wants possible visual directions;
- taste is ambiguous;
- an app/dashboard/landing page needs visual exploration before code;
- the workflow benefits from comparing 2-3 directions.
Default to three concept prompts:
- Conservative — familiar, low-risk, easy to implement.
- Strong-fit — best interpretation of the product brief.
- Divergent — more distinctive, useful for discovering taste boundaries.
Each prompt should specify:
- product category and target user;
- screen type and platform;
- information hierarchy;
- density and interaction posture;
- visual style and theme;
- what should not appear;
- that the result is a UI concept screenshot, not marketing art.
Image Prompt Template
Create a high-fidelity UI concept screenshot for . Platform: . Screen: . Target user: . Product feel: . Layout: . Must show: . Avoid: generic SaaS filler, fake irrelevant metrics, stock photos, excessive gradients, unreadable tiny text. Style: . The image should be a realistic product UI mockup suitable for turning into a design brief.
After Images Are Generated
If image artifacts are generated:
- Save or reference each image path/URL in the UI brief.
- Use vision analysis or manual review to extract what works and what does not.
- Update
Selected Directionwith the chosen layout, components, and visual rules. - Do not ask the build agent to copy the image blindly. Convert the image into text requirements, components, tokens, and acceptance criteria.
Post-Build Redesign Loop
After Superpowers or another coding agent builds the app, this skill can run again as a UI polish loop:
- Capture screenshots of the built app.
- Compare screenshots to the UI brief and product design doc.
- Identify UI gaps: hierarchy, spacing, navigation, copy, state handling, responsiveness.
- Optionally generate alternate UI concepts.
- Produce a focused redesign brief and implementation task list.
- Hand the task list to Superpowers with the constraint: improve UI only, avoid backend behavior changes unless explicitly approved.
Integration Rules
- This stage is Full-mode only by default.
- It must not make the workflow mandatory or brittle.
- If skipped, continue with the old design-doc → implementation-spec → handoff sequence.
- If the user wants speed, fold only a short
UI / UX Notessection into the design doc instead. - If image generation is unavailable, write image prompts and continue with a text-only UI brief.
- If an existing repo exists, inspect current UI/components before inventing a new design direction.
Common Pitfalls
- Letting images become the spec. Images are inspiration. The written UI brief and implementation handoff are the source of truth.
- Overcomplicating Lite mode. Do not add this stage to quick idea capture unless requested.
- Generic dashboard slop. Avoid fake metrics, irrelevant charts, and decorative cards unless the product actually needs them.
- Skipping states. Empty, loading, error, success, first-run, and mobile states matter for build quality.
- Breaking artifact numbering. If adding this stage to a package, update README artifact maps and subsequent stage numbers consistently.
- Copying proprietary references. Transform style principles; do not clone distinctive third-party UI.
Verification Checklist
- [ ] The workflow is Full mode or the user explicitly requested UI design.
- [ ] The product/design direction is clear enough for screen-level UI decisions.
- [ ]
02-ui-design-brief.mdexists if this stage was run. - [ ] Optional image prompts are stored in the brief even if image generation is skipped.
- [ ] Generated images, if any, are treated as references and translated into text requirements.
- [ ] Implementation handoff references the selected UI direction or explicitly states that UI is an assumption.
- [ ] If the stage is skipped, the old workflow still proceeds without blockage.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: AkoliteZA
- Source: AkoliteZA/hermes-agent-idea-workflow
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.