roadmap-with-validation

star 1

Interactive scoping, roadmap generation, and multi-agent validation

matthewod11-stack By matthewod11-stack schedule Updated 2/10/2026

name: roadmap-with-validation description: Interactive scoping, roadmap generation, and multi-agent validation disable-model-invocation: true allowed-tools: - Read - Write - Edit - Glob - Grep - Bash

Validation Protocol

Overview

Roadmap validation stress-tests the implementation plan before execution begins. Multiple validators review the roadmap looking for:

  • Missing dependencies
  • Underestimated complexity
  • Scope creep
  • Sequencing mistakes
  • Unclear acceptance criteria

Validation Checklist

1. Success Criteria Alignment

Question: Does every V1 feature directly contribute to the success criteria?

For each feature:

  • How does this help achieve the goal?
  • What happens if we cut this?
  • Is this actually V1, or V2 in disguise?

Red Flags:

  • Features justified as "nice to have"
  • Features that enable future features but don't serve v1
  • Polish items disguised as core functionality

2. Dependency Validation

Question: Is the sequencing actually correct?

Tests:

  • Can Phase 0 deliverables support Phase 1 work?
  • Are there hidden dependencies between "parallel" features?
  • Does any task assume something not explicitly built earlier?

Trace Test:

For each task:
  Task X requires -> [list everything] -> verify each exists earlier

Red Flags:

  • Tasks that "just need" something from another domain
  • Shared state modified by multiple features
  • Integration points not explicitly defined

3. Complexity Audit

Question: Are any tasks hiding significant complexity?

For each task:

  • Is this one task or 3-5 bundled together?
  • Does "implement X" hide research, design, iteration?
  • Are there edge cases not in acceptance criteria?

Red Flags:

  • Vague scope ("handle edge cases")
  • Tasks touching multiple files/domains
  • External API tasks
  • Anything called "simple" or "just"

4. Acceptance Criteria Quality

Question: Could someone verify each task without asking you?

For each criterion:

  • Specific and measurable?
  • Covers happy path AND error cases?
  • Verification method actually doable?

Red Flags:

  • "Works correctly" (what is correct?)
  • "Handles errors gracefully" (which errors? how?)
  • "User can do X" (under what conditions?)

Fix Pattern:

Before: "User can import recipes"
After: "User pastes URL -> system extracts title/ingredients/instructions ->
        displays preview -> user confirms -> recipe appears in list within 3 seconds.
        Error shown if URL invalid or parsing fails."

5. Risk Assessment

Categories:

  • Technical risk: Unproven tech, complex integrations
  • Scope risk: Features likely to expand
  • Dependency risk: External services, APIs
  • Knowledge risk: Learning as you build

For each risk:

  • Mitigation strategy
  • Fallback if it doesn't work
  • Phase where addressed

6. Parallelization Reality Check

If PARALLEL-READY:

  • Are agent boundaries truly independent?
  • What happens when both need shared files?
  • Is "read-only" list actually read-only?
  • How do agents communicate completion?

Common Issues:

  • Both domains need same component
  • Shared types evolve differently
  • Integration testing needs both domains

7. Out of Scope Validation

Questions:

  • All spec features appear in V1 or Out of Scope?
  • Any "out of scope" items needed for success criteria?
  • Can you resist adding these during execution?

Validator Prompt Template

# Roadmap Validation — [MODEL_NAME]

You are a skeptical technical reviewer. Your job is to stress-test this roadmap and find gaps before execution begins.

**Stance:** Assume something is wrong.

**Success Criteria:** [V1_SUCCESS_CRITERIA]

**Review the roadmap below and check:**
1. Success criteria alignment
2. Dependency validation
3. Complexity audit
4. Acceptance criteria quality
5. Risk assessment
6. Parallelization reality check (if applicable)
7. Out of scope validation

**Output Format:**

## Summary Verdict
[APPROVED / APPROVED WITH CHANGES / NEEDS REVISION]

## Key Findings
1. [Finding with severity: Critical/Important/Minor]
2. [Finding]
3. [Finding]

## Required Changes
- [ ] [Specific change needed]
- [ ] [Specific change needed]

## Risks Identified
| Risk | Severity | Mitigation | Phase |
|------|----------|------------|-------|

## Detailed Review
[Section-by-section analysis]

---

**ROADMAP TO REVIEW:**

[PASTE ROADMAP.md CONTENT]

Validation Verdicts

APPROVED

Roadmap is solid. No blocking issues found.

Criteria:

  • No critical findings
  • At most 1-2 minor suggestions
  • Dependencies validated
  • Acceptance criteria clear

APPROVED WITH CHANGES

Roadmap is workable but needs specific fixes.

Criteria:

  • No critical findings
  • 1-3 important findings with clear fixes
  • Can be fixed without restructuring
  • Fixes documented as checklist

NEEDS REVISION

Roadmap has fundamental issues requiring rework.

Criteria:

  • Critical findings (missing core dependencies, wrong sequencing)
  • Multiple important findings
  • Requires restructuring phases
  • Success criteria misalignment

Change Application

Change Categories

Critical (Must Fix):

  • Missing dependencies
  • Wrong phase sequencing
  • Success criteria misalignment

Important (Should Fix):

  • Unclear acceptance criteria
  • Complex tasks needing splits
  • Missing risk mitigations

Suggestions (Nice to Have):

  • Additional pause points
  • Documentation improvements
  • Polish items

Application Strategies

Dependency Addition:

Before: Phase 1: User Profile
After:  Phase 0: Foundation
        - [ ] Authentication setup
        Phase 1: User Profile (requires Auth)

Task Splitting:

Before:
- [ ] Build Dashboard

After:
- [ ] Build Dashboard Layout
- [ ] Add Dashboard Widgets
- [ ] Implement Dashboard Data Fetching

Acceptance Criteria Clarification:

Before:
- [ ] Implement Recipe Import
  - Acceptance: User can import recipes

After:
- [ ] Implement Recipe Import
  - Scope: URL paste, parsing, preview, confirmation
  - Acceptance: User pastes URL -> system extracts title/ingredients/instructions
                -> displays preview -> user confirms -> recipe appears in list
                within 3 seconds
  - Error Handling: Show error if URL invalid or parsing fails
  - Verification: Manual test with 5 different recipe sites

Validation Header

After validation passes, add to ROADMAP.md:

> **Validated:** [DATE]
> **Validation Sources:** claude, gpt4, grok, gemini
> **Status:** APPROVED WITH CHANGES
> **Changes Applied:** [N] critical, [N] important

Skip Conditions

Skip validation if:

  • Project is small (< 2 weeks)
  • You've built similar before
  • Roadmap came from extensive prior planning
  • Using Lite workflow tier

Don't skip if:

  • New domain for you
  • Large or complex spec
  • Multiple interacting features
  • Planning parallel execution

Protocol version: 1.0 | Created: 2026-02-01

Install via CLI
npx skills add https://github.com/matthewod11-stack/claude-setup --skill roadmap-with-validation
Repository Details
star Stars 1
call_split Forks 0
navigation Branch main
article Path SKILL.md
More from Creator
matthewod11-stack
matthewod11-stack Explore all skills →