Menü
Iterative 3-phase workflow with peer review cycle for changes needing validation (15-30 min)
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
Implementation and troubleshooting agent - builds features and fixes bugs
DevOps expertise for deployment, CI/CD, infrastructure, and automation
Adversarial quality gate agent for code review - finds flaws before users do
Product management expertise for story creation, backlog management, validation, and release coordination
Security audits, vulnerability analysis, and security best practices enforcement
Full 9-phase workflow for complex features, epics, and security-critical changes (2-4 hours)
| name | atlas-iterative |
| description | Iterative 3-phase workflow with peer review cycle for changes needing validation (15-30 min) |
Perfect for:
Time estimate: 15-30 minutes (including review cycles)
Success criteria:
Phase 1: Make Change → Implement change
Phase 2: Peer Review (Cycle) → Review → Fix → Repeat until pass
Phase 3: Deploy → Test + deploy
Goal: Implement the change you know needs to happen.
Understand what needs changing
Make the change
Self-verify
Example 1: Button spacing
// Before
<View style={{ padding: 8 }}>
<Button />
</View>
// After (better spacing)
<View style={{ padding: 16 }}>
<Button />
</View>
Example 2: Extract helper function
// Before (validation inline)
const isValid = email.includes('@') && email.length > 5
// After (extracted)
const validateEmail = (email) => {
return email.includes('@') && email.length > 5
}
const isValid = validateEmail(email)
Goal: Get peer feedback, address issues, repeat until approved.
1. Submit for review
2. Receive feedback
3. Address feedback
4. Re-submit
5. Repeat until PASS
Self-review first
# Run quick checks
npm run typecheck
npm run lint
Submit for peer review
Receive feedback
Address feedback
Re-submit
Before submitting:
During review:
Review pass criteria:
Feedback 1: Missing edge case
// Review: "What if activities array is empty?"
// Before
const firstActivity = activities[0]
// After
const firstActivity = activities.length > 0 ? activities[0] : null
if (!firstActivity) return null
Feedback 2: Convention violation
// Review: "Should use store-specific method"
// Before
useAppStore.setState({ users: newUsers })
// After
useUserStore.getState().setUsers(newUsers)
Feedback 3: Accessibility
// Review: "Gray text violates accessibility rules"
// Before
<Text style={{ color: '#666666' }}>Secondary text</Text>
// After
<Text style={{ color: '#000000' }}>Secondary text</Text>
If the atlas-agent-peer-reviewer skill is available:
"Review my changes: [brief description]
Files changed:
- /path/to/file1.js
- /path/to/file2.js
What I changed:
[Explanation]
Please check for:
- Edge cases
- StackMap conventions
- Code quality
"
The peer-reviewer agent will provide structured feedback with a verdict:
Goal: Deploy the approved changes.
Final validation
npm run typecheck
npm test
Update PENDING_CHANGES.md
## Title: Improve button spacing for better UX
### Changes Made:
- Updated button padding from 8px to 16px
- Applied consistently across login and signup screens
- Peer reviewed and approved
Deploy to QUAL
./scripts/deploy.sh qual --all
Verify deployment
Escalate to Standard workflow if:
Escalate to Full workflow if:
"Escalating to Standard workflow. Found 4 files need changes and complex edge cases require planning."
Then restart from Phase 1 of Standard workflow.
Use case: Adjust spacing, alignment, sizing for better UX
Pattern:
Time: 15-20 minutes
Use case: Extract logic, improve code organization
Pattern:
Time: 20-25 minutes
Use case: Update animations, transitions, visual effects
Pattern:
Time: 15-25 minutes
"Change looks good to me, deploying immediately"
Problem: Purpose of Iterative is validation. Without review, use Quick workflow.
Solution: Complete the review cycle or use Quick workflow if validation not needed.
Reviewer: "Missing edge case"
You: "Looks fine to me, merging anyway"
Problem: Defeats purpose of peer review.
Solution: Address all blocking feedback or escalate to discuss with team.
Started: "Adjust button padding"
Now doing: "Adjust padding + refactor button component + add new props"
Problem: No longer iterative, too complex.
Solution: Escalate to Standard workflow or split into multiple tasks.
// File: src/components/ActivityCard.js
// Before
const styles = StyleSheet.create({
card: {
padding: 12,
margin: 8
},
title: {
fontSize: 16,
marginBottom: 4
}
})
// After (improved spacing)
const styles = StyleSheet.create({
card: {
padding: 16, // More breathing room
margin: 12 // Better separation between cards
},
title: {
fontSize: 18, // Larger, more prominent
marginBottom: 8 // Better separation from subtitle
}
})
Self-verify: Looks better visually ✅
Submit: "Updated card spacing for better hierarchy. Please review."
Feedback received:
Address feedback:
// Tested on Android simulator - widths still work ✅
// Tested on iPhone SE - spacing looks good ✅
Re-submit: "Tested on Android and small screens, all good."
Feedback received:
Update PENDING_CHANGES.md:
## Title: Improve activity card spacing for better visual hierarchy
### Changes Made:
- Increased card padding from 12px to 16px
- Increased card margin from 8px to 12px
- Increased title font size from 16px to 18px
- Increased title bottom margin from 4px to 8px
- Tested on Android and iOS small screens
- Peer reviewed and approved
Deploy:
./scripts/deploy.sh qual --all
# ✅ Deployed successfully
Total time: 20 minutes ✅
# Validation
npm run typecheck
npm run lint
# Deploy
./scripts/deploy.sh qual --all
The Iterative workflow adds peer validation to simple changes. Use it when:
Key advantage: Catches edge cases and convention violations early through structured review.
Remember: If review reveals complexity, escalate to Standard workflow rather than forcing it through Iterative.