validate-consistency

validate-consistency is a command for Claude Code from navraj007in/architecture-cowork-plugin. It costs 15 tokens per session (4,336 once invoked), scanned A, original, Apache-2.0.

A command that checks architecture and project documents for conflicts, missing links, outdated outputs, and invalid data formats. It reports the problems by severity and can fix some simple conflicts.

In plain words
What is it for?
Use it to validate architecture, scaffolding, design settings, project state, and other generated outputs for consistency.
Why use it?
It catches contradictions that can appear when many commands update shared project files over time.

Command for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the architect plugin — 48 skills, 63 commands, 19 agents, 7 MCP servers shipped together

Good fit Use it to validate architecture, scaffolding, design settings, project state, and other generated outputs for consistency.

Compare 6 commands from other repositories ↓
Install with agentmods
npx agentmods add commands/navraj007in/architecture-cowork-plugin/validate-consistency
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/navraj007in/architecture-cowork-plugin

Made for: Claude Code.

Or install architect, the plugin that ships this one along with the rest of its 48 skills, 63 commands, 19 agents, 7 MCP servers.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for validate-consistency

README.md
[![agentmods](https://agentmods.dev/badge/commands/navraj007in/architecture-cowork-plugin/validate-consistency/github.svg)](https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/validate-consistency)
Your own site
<a href="https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/validate-consistency"><img src="https://agentmods.dev/badge/commands/navraj007in/architecture-cowork-plugin/validate-consistency/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for validate-consistency

Your own site · 80×15
<a href="https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/validate-consistency"><img src="https://agentmods.dev/badge/commands/navraj007in/architecture-cowork-plugin/validate-consistency.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 15 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,336 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00015 $0.04336
Opus 5.5 $0.00006 $0.01734
Sonnet 5.5 $0.00003 $0.00867
Haiku 4.5 $0.00002 $0.00434

Measured 1mo ago against content hash ba00f216e047, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

validate-consistency scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

commands/validate-consistency.md · 464 lines

How it starts

The opening of the file, as written. The whole thing — 464 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/architect:validate-consistency

Scans all architecture-output files and detects conflicts between outputs from different commands (e.g., scaffold creates a component that blueprint says shouldn't exist, or design tokens contradict _state.json colors). Reports severity-categorized conflicts and suggests resolution order.

Trigger

/architect:validate-consistency
/architect:validate-consistency [--fix]      # auto-fix simple conflicts
/architect:validate-consistency [--detailed] # show conflict reasoning
/architect:validate-consistency [--rules <rule-set>]  # validate against specific rule set

Purpose

When dozens of commands write to the same architecture-output directory over weeks of work, inconsistencies accumulate:

  1. Drift: blueprint says component X has feature Y, but scaffold doesn't generate Y
  2. Contradictions: design-system sets primary color to #f97316, but _state.json.design.primary is #0ea5e9
  3. Stale outputs: cost-estimate ran 30 days ago with 8 components, but scaffold now has 12 components
  4. Schema violations: component in scaffold.json has port "3000" (string), but _state.json expects number
  5. Missing links: entities in _state.json reference fields that don't exist in data-model output

This command scans the entire state + architecture-output and reports all conflicts with clear remediation.

Input

No required input. Reads:

  • architecture-output/_state.json
  • architecture-output/_activity.jsonl
  • architecture-output/*.md (all generated docs)
  • architecture-output/contracts/ (if present)
  • architecture-output/design-system/ (if present)
  • Project source files (if scaffolded)

Output Files

1. architecture-output/consistency-report.md (always generated)

# Consistency Validation Report

**Generated:** 2026-04-24 14:32:10  
**Project:** my-startup  
**Scan scope:** 52 files analyzed, 8 conflicting findings

---

## Summary

✅ **Pass Rate:** 44/52 outputs consistent (85%)  
❌ **Critical Conflicts:** 1  
⚠️  **Warnings:** 3  
ℹ️  **Info:** 4

**Bottom Line:** One critical conflict that must be fixed before launch. Three warnings that should be addressed this sprint.

---

## Critical Conflicts

### CONF-001: Design Token Contradiction (design)

**Conflict:** `_state.json.design.primary` (#0ea5e9) ≠ `design-tokens.json.primary` (#f97316)  
**Scope:** Affects all UI components using primary color  
**Root cause:** design-system command ran on 2026-04-20 with one palette, but blueprint ran on 2026-04-24 with updated palette. design-tokens.json was not regenerated.  
**Impact:** Components generated after 2026-04-20 use old color (#0ea5e9); new scaffold will use blueprint color (#f97316); inconsistent UI.

**Remediation:**
1. Decide which color is correct: old (#0ea5e9) or new (#f97316)?
   - Option A: Keep new (#f97316) — aligns with latest blueprint
   - Option B: Keep old (#0ea5e9) — revert blueprint change
2. If keeping new: Run `/architect:design-system --force-regenerate` to update tokens
3. If keeping old: Run `/architect:blueprint` with old color, then regenerate design tokens
4. Verify: `jq '.design.primary' _state.json` matches `jq '.primary' design-system/design-tokens.json`

**Time to fix:** 10 minutes

---

## Warnings

### WARN-001: Stale cost-estimate (cost_estimate)

**Conflict:** `cost-estimate` last ran 2026-04-10 (14 days ago) with 8 components. Current scaffold has 12 components.  
**Scope:** Cost projections are outdated  
**Impact:** Budget estimates are 33% incomplete  

**Remediation:**
1. Run `/architect:cost-estimate --regenerate` to refresh estimates
2. Compare with old estimate and validate architecture hasn't changed significantly
3. Update budget if needed

**Time to fix:** 5 minutes

---

### WARN-002: Missing entity references in schema (entities)

**Conflict:** `_state.json.entities[]` references field `user.profile_url`, but `data-model/schema.prisma` does not define this field.  
**Scope:** One field missing from schema  
**Impact:** ORM schema and state mismatch; future scaffold might not include this field

**Remediation:**
1. Decide: is `profile_url` a real field that should be in schema?
   - Option A: Yes → Run `/architect:generate-data-model --add-field User.profile_url`
   - Option B: No → Remove from _state.json: `jq 'del(.entities[] | select(.name=="User") | .fields[] | select(.name=="profile_url"))'`
2. Regenerate schema: `jq '.entities' _state.json > entities-for-schema.json` and pass to data-model
3. Verify all entities match schema

**Time to fix:** 15 minutes

---

### WARN-003: Component port collision (components)

**Conflict:** Two components claim the same port: `api-server` (port 3000) and `worker-service` (port 3000 in fallback env)  
**Scope:** Port conflict when running locally  
**Impact:** `npm run dev` will fail with "EADDRINUSE: port 3000 already in use"

**Remediation:**
1. Check scaffold output for port assignments: `grep -r "PORT" . --include="*.env.example"`
2. Reassign: worker-service should use port 3001 (or configurable)
3. Update _state.json: `jq '.components[] | select(.name=="worker-service").port = 3001'`
4. Verify no more collisions: `jq '.components[].port' _state.json | sort | uniq -d`

**Time to fix:** 10 minutes

---

### WARN-004: Tech stack version mismatch (tech_stack)

**Conflict:** `_state.json.tech_stack.backend` lists "Node.js 18", but scaffold `package.json` requires "Node.js ^20"  
**Scope:** Node.js version specification  
**Impact:** Developers using Node 18 will see build errors; CI/CD might use wrong node version

**Remediation:**
1. Update `package.json` in scaffold to match state or vice versa
2. Option A: Update _state.json to "Node.js ^20"
3. Option B: Update package.json to `"engines": { "node": ">=18" }`
4. Choose one source of truth and stick with it

**Time to fix:** 5 minutes

---

## Info (Suggestions, Not Blocking)

### INFO-001: Design personality unused (design)

**Info:** `_state.json.design.personality` is "bold-commercial", but scaffold doesn't use this personality in component naming/structure.  
**Suggestion:** If personality is intentional, add it to component metadata for design consistency

---

### INFO-002: Market research outdated (market_research)

**Info:** `market_research` field in _state.json was last updated 2026-04-15 (9 days ago).  
**Suggestion:** Consider refreshing market research every 2 weeks; rerun `/architect:deep-research` if market has shifted

---

### INFO-003: Multiple design system outputs

**Info:** Both `design-system/design-tokens.json` and `design-tokens-legacy.json` exist.  
**Suggestion:** Remove legacy file to reduce confusion: `rm design-tokens-legacy.json`

---

### INFO-004: No monitoring setup yet

**Info:** Scaffold complete, but `/architect:setup-monitoring` hasn't run yet.  
**Suggestion:** Add monitoring before launch; currently no observability in place

---

## Conflict Categories

All conflicts fit one of these categories:

| Category | Example | Severity | Auto-fixable |
|----------|---------|----------|--------------|
| **State contradictions** | design.primary in _state.json ≠ design-tokens.json.primary | Critical | Partial (choose source) |
| **Stale outputs** | cost-estimate is 30 days old | Warning | No (requires rerun) |
| **Schema violations** | component.port is string "3000" not number | Critical | Yes (type coercion) |
| **Missing references** | entity references field not in schema | Warning | No (requires investigation) |
| **Duplicates** | Two components with same name | Critical | No (requires choice) |
| **Naming inconsistencies** | Component slug in scaffold ≠ _state.json name | Warning | Partial (normalize) |
| **Version mismatches** | Tech stack version ≠ package.json engines | Warning | No (choose source) |
| **Unused outputs** | Old tokens file exists alongside new one | Info | Yes (delete) |

---

## Rules Applied

Load `skills/consistency-rules/SKILL.md` for the full rule catalog and category definitions before applying Steps 3–5 below.

Validation uses 23 rules (see `/architect:validate-consistency --detailed` for full list):

**State Rules (6):**
- RULE-S-001: Design colors must be valid hex (#RRGGBB)
- RULE-S-002: Component ports must be unique
- RULE-S-003: Entity names must match schema (if schema exists)
- RULE-S-004: Tech stack versions must parse as valid semver
- RULE-S-005: Persona IDs must be unique
- RULE-S-006: All referenced entities must exist

**Output Rules (8):**
- RULE-O-001: Design tokens file must match _state.json colors
- RULE-O-002: Scaffold files must match component names in _state.json
- RULE-O-003: Cost estimate must be based on current component count
- RULE-O-004: Test coverage % must not exceed 100%
- RULE-O-005: Compliance rules must reference existing entities
- RULE-O-006: Monitoring metrics must align with tech stack
- RULE-O-007: Load test scenarios must use valid endpoints
- RULE-O-008: Documentation must reference existing components

**Cross-Command Rules (9):**
- RULE-X-001: No component should exist in both removed and added lists
- RULE-X-002: Design personality must be consistent across scaffold + design-system
- RULE-X-003: Entity count in _state.json ≥ entity count in schema (or equal)
- RULE-X-004: Blueprint architecture must match scaffold service layout
- RULE-X-005: Tech stack languages must match codebase (.ts/.py/.go/etc.)
- RULE-X-006: Monitoring observability provider must be listed in tech_stack.integrations
- RULE-X-007: Compliance frameworks must be supported by chosen tech stack
- RULE-X-008: Load test target RPS must be realistic for tech stack
- RULE-X-009: All referenced external services must exist in tech_stack.integrations

---

## Conflict Resolution Order

When multiple conflicts exist, fix in this order to avoid cascading issues:

1. **Critical state contradictions** (CONF-001, CONF-002, etc.)
   - Impact: Foundation of entire project; everything else depends on consistent state
   - Example: design color mismatch
   - Fix time: 5-15 min each

2. **Critical schema violations** (CONF-003, etc.)
   - Impact: Breaks type safety and code generation
   - Example: component.port is string not number
   - Fix time: 3-10 min each

3. **Warnings about missing outputs** (WARN-001, WARN-002, etc.)
   - Impact: Incomplete architecture; might miss important design decisions
   - Example: stale cost-estimate; missing entity in schema
   - Fix time: 5-20 min each

4. **Warnings about mismatches** (WARN-003, WARN-004, etc.)
   - Impact: Local dev/deploy might fail; CI/CD might use wrong config
   - Example: port collisions; version mismatches
   - Fix time: 5-15 min each

5. **Info suggestions** (INFO-001, INFO-002, etc.)
   - Impact: Nice-to-have; doesn't block anything
   - Example: outdated market research; unused design personality
   - Fix time: 1-5 min each

---

## Auto-Fix Behavior (`--fix` flag)

When `--fix` is enabled, load `skills/conflict-resolution/SKILL.md` for the step-by-step remediation procedure matching each rule ID before applying a fix — it defines root cause and safe-fix boundaries per rule.

When `--fix` is enabled, only safe conflicts are auto-fixed:

✅ **Auto-fixed:**
- Type coercion (string port → number)
- Removing duplicate files (`design-tokens-legacy.json`)
- Normalizing component names (kebab-case consistency)
- Adding missing _state_version field
- Fixing color format (add # prefix if missing)

❌ **Not auto-fixed (requires user choice):**
- Design color contradictions (which color is correct?)
- Port collisions (which service keeps the port?)
- Missing entities in schema (add or remove field?)
- Stale outputs (rerun or accept old data?)
- Version mismatches (upgrade tech stack or revert?)

For non-auto-fixable conflicts, `--fix` shows remediation steps and stops, waiting for user to resolve.

---

## Detailed Mode (`--detailed` flag)

Shows extended reasoning for each conflict:

```markdown

Read the full file on GitHub · 464 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 1mo ago First seen · 464 lines · 0 tokens per session scan A ba00f216e047

Subscribe to this mod's changes

validate-consistency is a command published in the GitHub repository navraj007in/architecture-cowork-plugin (2 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 15 tokens to every session and 4,336 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.