ux-flows
Skill187 starsChanged 5 months agoMITView on GitHub
Use to derive the user flows of a UI-touching feature after the spec is clarified — one mermaid flowchart per UI-touching §4 user story (happy + alt/error branches from §5 ACs), a screen inventory (SCR-NN ids), and an AC→flow map, written to docs/features/{slug}/ux-flows.md. Triggers on "ux flows for {slug}", "user flows for {slug}", "screen flow for {slug}", "/sdd:ux-flows {slug}", "юзер-флоу для {slug}", "потік екранів {slug}", "намалюй флоу користувача". Always markdown + mermaid regardless of the design tool; the Socratic pass confirms each flow in prose, never raw mermaid. Feeds design (target-surface evidence), sequences (SCR alignment), screens (the inventory) and plan-tests (e2e-through-UI paths). Hard-refuse if spec.md is missing; skipped for features with no human-facing UI.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- JoseArron/skillstheir own version, first seen Oct 11, 2026
- djalmajr/skillstheir own version, first seen Sep 4, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- djalmajr/skills★ 4same name
- JoseArron/skills★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository22 other files it gives its agents.
- apiSkill
- clarifySkill
- classify-sizeSkill
- configSkill
- data-modelSkill
- decide-adrSkill
- designSkill
- design-systemSkill
- fixSkill
- glossarySkill
- implementSkill
- interviewSkill
- plan-testsSkill
- reviewSkill
- roadmapSkill
- scaffoldSkill
- screensSkill
- sequencesSkill
- shipSkill
- specifySkill
- surveySkill
- tasksSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 63
Set the depth dial + platform: Read interview_depth from .claude/sdd.local.md (else medium).
- Distinctive: line 104
Every flow was confirmed in prose (or written + ledgered at easy).
- Distinctive: line 105
No visual design leaked in: no component names, no layout, no styling — flow altitude only (screens is where states + components live).
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skill: ux-flows
Draws how the user moves through a UI-touching feature — after clarify, before design. For
each UI-touching §4 user story it produces a mermaid flowchart (happy path + the alt/error
branches the §5 ACs demand), builds the screen inventory (SCR-NN — the id contract screens
details later), and maps every UI-touching AC to the flow/branch that shows it. The artifact is
always markdown + mermaid whatever docs/design-system.md picks as the drawing tool — flows
are flow-altitude, not visual design. design then reads it as evidence for the
target-surface + UI-architecture decisions (the formal target_surfaces declaration stays
design's); sequences aligns UI-driven flows on the SCR ids; plan-tests takes the
e2e-through-UI paths from here.
This stage is optional by surface, not by size: its N/A condition (no human-facing UI) lives in
../_shared/size-matrix.md and is evaluated by clarify's handoff
(specify's when clarify was legally skipped). Invoked directly, it always runs.
Question phrasing → ../_shared/ask-style.md; each diagram is confirmed
in prose, never as raw mermaid → ../_shared/diagram-presentation.md.
Flow labels + prose follow artifact_language — mermaid keywords, SCR ids and AC ids stay English
→ ../_shared/artifact-language.md.
Owner
PM + designer (or whoever owns the user experience). The PM confirms each flow matches a real user
story; the Tech Lead flags flows that imply architecture (they become design input, not decisions
here).
Inputs
<slug>— feature slug.- Gate (hard-refuse if missing):
docs/features/<slug>/spec.md— the flows derive from §4 user stories + §5 ACs. Absent → STOP: «runspecify <slug>first — ux-flows derives from its user stories». - (Expected)
docs/design-system.md— the platform posture (the default platform assumption) + the tool. Absent → not a block: work incode-mode assumptions, recommend/sdd:design-systemin the handoff. - (Optional)
CONTEXT.md(both levels, per-feature wins) — canonical roles for the actors. - (Optional)
docs/features/<slug>/.size/.route— depth + handoff resolution; absent → default M / standard and say so in the handoff.
Protocol
- Gate + read.
test -f docs/features/<slug>/spec.md→ missing = refuse with the pointer above. Read spec §1 (context), §4 (user stories — which touch a UI?), §5 (ACs),CONTEXT.mdglossary, anddocs/design-system.md(posture + tool; note its absence for the handoff). - Set the depth dial + platform. Read
interview_depthfrom.claude/sdd.local.md(else medium); unless--depth=was passed, ask ONE depth-selectionAskUserQuestionper../_shared/ask-style.md, then confirm the platform posture in the same call (second question): the design-system posture as «(Recommended)», deviation allowed + recorded with its why. Depth governs the per-flow question volume (→../_shared/interview-depth.md); coverage never shrinks. - Derive flows + inventory. For every UI-touching §4 user story: one flow — happy path +
an alt/error branch per relevant §5 AC. Collect every screen the flows visit into the
Screen inventory (
SCR-NN+ purpose/entry/exit); flow nodes reference the SCR ids. A backend-only user story is listed as out of scope, not drawn. - Socratic pass — prose, never raw mermaid. Per
../_shared/diagram-presentation.md: write each flow intodocs/features/<slug>/ux-flows.md(bootstrapped from./templates/ux-flows.mdon the first flow; each flow lands on disk as soon as it's drafted — progressive writes per../_shared/write-and-commit.md), validate it parses per../_shared/mermaid-check.md, then describe it in plain words (every branch) and confirm — at medium/hard oneAskUserQuestionper flow (Accept / Fix / Save-as-OQ / Drop); at easy, write + one-line summary into the assumptions ledger and proceed. - Fill the AC map + write. Complete the AC-coverage table (every UI-touching §5 AC →
flow/node/branch, or an explicit
N/A: <reason>) on disk, re-validate every mermaid block, stampupdated_at. No commit yet — it follows the self-check. - Structural self-check — per
../_shared/self-check.md: re-read the file from disk and verify 4 items: (1) every UI-touching §4 user story has a flow; (2) every flow node's SCR id exists in the inventory (and every inventory row appears in ≥1 flow); (3) every UI-touching §5 AC appears in the coverage table with a flow/branch or an explicit N/A; (4) every mermaid block parses. Fix + re-check ≤2 cycles; surface anything unresolved. - Commit + handoff. On a passing self-check, commit per
../_shared/write-and-commit.md— one commit forux-flows.md(ux-flows: <slug> ux-flows.md); an unresolved item means no commit. Then emit the stage-handoff block per../_shared/handoff.md— What I did (incl. «self-check: 4/4 pass» + the commit SHA; + «docs/design-system.mdabsent — run/sdd:design-system» when it was) + Review (docs/features/<slug>/ux-flows.md) + Run next:/clear, then/sdd:design <slug>(it reads these flows as target-surface evidence).
Definition of Done
docs/features/<slug>/ux-flows.mdexists: platform decisions, the SCR-NN inventory, one flow per UI-touching §4 user story (happy + AC-demanded branches), the AC-coverage table — zero silently uncovered UI ACs.- Every flow was confirmed in prose (or written + ledgered at easy); every mermaid block parses.
- No visual design leaked in: no component names, no layout, no styling — flow altitude only (screens is where states + components live).
Anti-patterns
- Drawing screens here. Components, states, layout belong to
screens; this artifact is the movement between screens, not their content. - Deciding architecture here. «SPA vs SSR», «this needs a websocket» — flag it as design input;
designdecides and declarestarget_surfaces. - Raw mermaid as the confirmation prompt — the anti-pattern
diagram-presentation.mdexists to kill. - Skipping backend-only stories silently. List them as out of scope with one line — the reader must see they were considered.
- Blocking on a missing design-system. Its absence degrades (code-mode assumptions + a handoff recommendation), never blocks the flow work.
References & template
./templates/ux-flows.md— output scaffold; inline comments are the generation contract.../_shared/diagram-presentation.md·../_shared/mermaid-check.md— confirm-in-prose + parse-validation.../_shared/size-matrix.md— the N/A condition (no human-facing UI) evaluated by the upstream handoff.../_shared/interview-depth.md— the depth dial set in step 2.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

