Install
$ agentstack add skill-jakeintech-claude-ios-skills-ios-prd ✓ 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
iOS Product Requirements Document — Hive Mind
Turn a raw app idea into a complete, buildable PRD using 5 specialist analyst agents working in parallel.
Input
Raw app idea via $ARGUMENTS. Can be anything from a sentence to a paragraph:
- "Golden Hour: sunrise/sunset widgets, daily photo capture, metadata tracking, monthly recaps, community submissions"
- "A habit tracker that uses HealthKit to correlate habits with health metrics"
- "Minimalist podcast player focused on walking commutes"
Phase 1: Parse & Dispatch
1.1 Parse the Idea
Read the raw idea and create a structured brief:
- Core concept: one sentence distillation
- Key features mentioned: bulleted list
- Implied features: what the user probably wants but didn't say
- Target audience: who is this for
1.2 Dispatch 5 Analyst Agents
Dispatch ALL 5 agents in parallel using the Agent tool. Each agent receives:
- The raw idea
- The structured brief from 1.1
- Their specific analyst prompt from
${CLAUDE_SKILL_DIR}/agents/{analyst}.md - The patterns reference:
${CLAUDE_SKILL_DIR}/patterns.md - The frameworks reference:
${CLAUDE_SKILL_DIR}/frameworks.md
Read each agent prompt file before dispatching:
Agent 1: Product Strategist
prompt file: ${CLAUDE_SKILL_DIR}/agents/product-strategist.md
model: default (needs web search for competitive analysis)
Agent 2: Platform Engineer
prompt file: ${CLAUDE_SKILL_DIR}/agents/platform-engineer.md
model: default (needs context7 for Apple docs)
Agent 3: Design Analyst
prompt file: ${CLAUDE_SKILL_DIR}/agents/design-analyst.md
model: sonnet (screen inventory is mechanical)
Agent 4: Data Architect
prompt file: ${CLAUDE_SKILL_DIR}/agents/data-architect.md
model: default (needs context7 for SwiftData docs)
Agent 5: Growth & Compliance
prompt file: ${CLAUDE_SKILL_DIR}/agents/growth-compliance.md
model: default (needs web search for App Store guidelines)
Wait for all 5 to complete.
Phase 2: Synthesis
Merge all 5 analyst outputs into a unified draft PRD:
- Deduplicate — multiple agents identify the same feature/integration → merge, keep the most detailed version
- Cross-reference — verify consistency:
- Every framework the Platform Engineer recommends has permissions from Growth & Compliance
- Every entity the Data Architect defines supports features from the Platform Engineer
- Every screen the Design Analyst proposes is technically feasible per Platform Engineer
- Nothing violates the Product Strategist's hard boundaries
- Resolve conflicts — if agents disagree:
- Technical feasibility wins over design wishes
- Hard boundaries win over feature suggestions
- Compliance requirements are non-negotiable
- Document the conflict and resolution
- Produce unified draft PRD with sections A through F (see below)
Phase 3: Cross-Review
Dispatch a second round — each analyst reviews the merged PRD for their domain:
Dispatch 5 review agents in parallel, each receiving the merged PRD:
- Product Strategist: "Does anything violate the boundaries? Are epics correctly prioritized?"
- Platform Engineer: "Are all features technically feasible? Any missing frameworks?"
- Design Analyst: "Is the screen inventory complete for all features?"
- Data Architect: "Does the schema support all features? Any missing relationships?"
- Growth & Compliance: "Does every framework have proper permissions? Any App Store risks?"
Incorporate review feedback. If no issues, finalize.
Phase 4: Present to User
Present the complete PRD in these sections:
Section A: Product Vision
- North Star statement
- Hard boundaries (what this app is NOT) — 5-8 items
- Soft boundaries (where to stretch carefully)
- Pillars (2-4 core domains)
- Design identity (color mood, typography feel, icon style, tone)
- Competitive landscape (from Product Strategist's blue ocean analysis)
- Blue ocean opportunities (what competitors all miss)
Section B: Technical Architecture
- Framework matrix table: Framework | API | Info.plist | Entitlement | Privacy | iOS Version
- Platform integrations — split into:
- Core features (what user asked for)
- Autonomously identified integrations (with reasoning)
- Considered & rejected (with reasoning)
Section C: Data Model
- Entity diagram (ASCII or Mermaid)
- SwiftData
@Modelschema overview - Persistence strategy (on-device, CloudKit, or hybrid)
- Sync strategy if applicable
Section D: Epic Breakdown
- Numbered epics with build order
- Each tagged with: skill handoff, frameworks, complexity (low/medium/high)
Section E: Compliance
- All permissions with usage description strings (from permissions.md patterns)
- Privacy manifest declarations
- Age rating assessment
- Export compliance determination
Section F: Monetization & Growth
- Pricing model (free, freemium, subscription)
- Feature gating — which features are free vs premium
- Growth loops (organic virality, share moments, widget exposure)
Ask user to review and approve.
Phase 5: Handoff
On approval, automatically flow through the next steps — no manual invocation needed.
5.1 Generate PRD Documents
Save these files in a temporary staging area (current directory):
docs/product-vision/00-product-bounds.md— pre-filled from Section Adocs/product-vision/SDLC-WORKFLOW.md— pre-filled with epics from Section Ddocs/product-vision/data-model-spec.md— Section C for ios-data-model to consumedocs/product-vision/widget-spec.md— widget requirements for ios-widget (if applicable)docs/product-vision/framework-requirements.md— Section B + E for ios-scaffold to consume
5.2 Auto-invoke /ios-scaffold
Extract the app name from the idea (use the North Star or ask user to confirm the name). Then invoke the ios-scaffold skill:
Skill: ios-scaffold
Args: {AppName}
This creates the full project directory with the PRD documents moved into place.
5.3 Auto-invoke domain skills
After scaffold completes, automatically invoke the next skills in sequence:
- If data model was specified: invoke
/ios-data-model— it readsdocs/product-vision/data-model-spec.mdand generates the SwiftData schema - If widgets were identified: invoke
/ios-widgetwith the widget description from the PRD - If CI was discussed: suggest
/ios-cibut don't auto-invoke (one-time setup, user may want to configure)
5.4 Report
After all auto-invocations complete, report:
- Project created at
./{AppName}/ - What was auto-generated (scaffold, data model, widgets)
- Epic build order — what to work on first
- Next manual steps: "Start building Epic 1. Use
/ios-iterateand/ios-code-reviewas you go." - Ship checklist: "When ready, run
/ios-app-icon→/ios-screenshots→/ios-store-listing→/ios-privacy→/ios-testflight→/ios-submit"
The goal: user types one command with their idea, and walks away with a fully scaffolded project including data models and widgets ready to build on.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Jakeintech
- Source: Jakeintech/claude-ios-skills
- 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.