Help

Vladyslav Skills — Help

Type: Engineer (light)

One-Terminal Workflow (v3.x)

Run any skill from a single Opus session. No manual /model switching required.

• Architect skills run interactively in Opus main session. Internal Agent(...) dispatches annotated with model="sonnet" (executor work) or model="opus" (synthesis/research).
• Engineer (light) — bash-driven skills run pre-flight Q&A in Opus main, then a bash script does the work, then a summary is rendered.
• Engineer (light) — Opus inline skills (write-user-stories, write-test-docs, write-project-docs, compact-save, help) run pre-flight Q&A + LLM-driven generation entirely in Opus main, with no subagent dispatch.

Available Skills (/vladyslav:)

Architect:

| Skill | Purpose | |-------|---------| | ingest | Existing-project intake: architecture docs + MemPalace seed in one scan pass | | add-feature | Add feature (full cycle, 9 superpowers) | | fix-bug | Fix bug (full cycle, 7 superpowers) | | discover | Auto-fill product/start-project.md via AI research | | discover-apple-check | Apple App Store compliance pre-check (iOS only) | | design-sync | Extract design tokens from code into docs/design/system.md | | design-page | Design app screens in Pencil via parallel subagents | | swiftui-pro | SwiftUI/Swift code review (iOS 26 / Swift 6.2) |

Engineer (light) — bash-driven:

| Skill | Purpose | |-------|---------| | init-project | Create new project — bare AI shell (minimal) or interactive module menu (scripts/modules/*) | | attach-project | Add structure to existing project (calls scripts/attach-project.sh) | | pre-release-check | Pre-release verification (calls scripts/pre-release-checks.sh) |

Engineer (light) — Opus inline:

| Skill | Purpose | |-------|---------| | write-user-stories | Update user stories | | write-test-docs | Test plan + manual QA docs | | write-project-docs | README, onboarding, deployment | | compact-save | Snapshot task state to MemPalace (auto before compact) | | help | This reference |

Recommended Workflows

New project:

init-project → discover → add-feature → write-test-docs → pre-release-check

Existing project:

attach-project → ingest → add-feature

Before release:

write-user-stories → write-test-docs → write-project-docs → pre-release-check

Bug fix:

fix-bug → write-test-docs → pre-release-check

Superpowers Integration

All 13 non-meta superpowers skills are integrated:

| Superpowers Skill | Used In | When | |-------------------|---------|------| | brainstorming | add-feature | Design phase | | writing-plans | add-feature | Planning phase | | executing-plans | add-feature | Execution (parallel session) | | subagent-driven-development | add-feature | Execution (this session) | | dispatching-parallel-agents | add-feature, ingest | Parallel components | | using-git-worktrees | add-feature, fix-bug | Isolated branch | | test-driven-development | add-feature, fix-bug, write-test-docs | Tests + implementation | | systematic-debugging | fix-bug | Diagnose root cause | | requesting-code-review | add-feature, fix-bug | After implementation | | receiving-code-review | add-feature, fix-bug | Process feedback | | finishing-a-development-branch | add-feature, fix-bug | Merge/PR | | verification-before-completion | pre-release-check | Evidence-based checks | | writing-skills | (meta) | Edit vladyslav skills |

Rules

• Model control: Run from Opus session. Skills handle Sonnet dispatch internally — no manual /model switches needed.
• Translations: Do NOT add until pre-release-check phase.
• Commits: No AI mentions (no Co-Authored-By, no "Generated by Claude").
• Internal Agent dispatches: Architect skills annotate every Agent(...) call with explicit model= ("sonnet" for executor work, "opus" for synthesis/research).
• Superpowers dependency: Make sure superpowers plugin is installed.

Stack Options

Backend (single choice): python (default) / go / other / none Frontend/Mobile (multi-select): flutter / swift / kotlin / other / none

All stacks: CORE (bare AI shell) is always written. Docs / backend-infra / agents are opt-in via the interactive menu. minimal mode = CORE only.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: VladislavSournine
• Source: VladislavSournine/vladyslav-skills
• License: MIT

Install and usage instructions live in the source repository linked above.