architecture-judge

architecture-judge is an agent for Claude Code from gonzalezpazmonica/savia. It costs 17 tokens per session (698 once invoked), scanned A, original, MIT.

A code-review judge focused on software architecture. It checks whether code respects layer boundaries, avoids excessive coupling, follows project patterns, and keeps concerns separate.

In plain words
What is it for?
Use it to review dependencies between layers, circular relationships, overly large classes, unnecessary abstractions, and mixed responsibilities such as SQL or HTML in the wrong place.
Why use it?
Code can work while still making a project harder to change, such as when a controller talks directly to a database layer. This judge identifies those structural problems and over-engineering.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions OpenCode.

Part of the pm-workspace plugin — 124 commands, 75 agents shipped together

Good fit Use it to review dependencies between layers, circular relationships, overly large classes, unnecessary abstractions, and mixed responsibilities such as SQL or HTML in the wrong place.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/gonzalezpazmonica/savia/architecture-judge
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/gonzalezpazmonica/savia

Made for: Claude Code.

Or install pm-workspace, the plugin that ships this one along with the rest of its 124 commands, 75 agents.

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 architecture-judge

README.md
[![agentmods](https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/architecture-judge/github.svg)](https://agentmods.dev/agents/gonzalezpazmonica/savia/architecture-judge)
Your own site
<a href="https://agentmods.dev/agents/gonzalezpazmonica/savia/architecture-judge"><img src="https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/architecture-judge/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 architecture-judge

Your own site · 80×15
<a href="https://agentmods.dev/agents/gonzalezpazmonica/savia/architecture-judge"><img src="https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/architecture-judge.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 17 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 698 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 $0.00017 $0.00698
Opus 5 $0.00009 $0.00349
Sonnet 5 $0.00003 $0.00140
Haiku 4.5 $0.00002 $0.00070

Measured 5d ago against content hash 1d98932aeafe, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

architecture-judge 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 5d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

.claude/agents/architecture-judge.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Architecture Judge

You are one of 5 judges in the Code Review Court. Your focus: architectural integrity.

What you check

  1. Layer violations: controller calling repository directly, domain importing infrastructure
  2. Coupling: fan-out > 3 services from one file, circular dependencies, god classes
  3. Abstraction level: single-use factories, unnecessary indirection, premature abstraction
  4. Pattern consistency: does the code follow the project's established patterns?
  5. Separation of concerns: HTML in backend, SQL as strings, CSS in logic (per template-separation.md)
  6. Over-engineering: unnecessary abstractions that AI reliably introduces

What you DON'T check

  • Logic correctness → correctness-judge
  • Security → security-judge
  • Naming/complexity → cognitive-judge
  • Spec compliance → spec-judge

Output format (YAML)

judge: "architecture-judge"
reviewed_at: "{ISO timestamp}"
files_reviewed: ["{file1}", "{file2}"]
verdict: "pass|conditional|fail"
findings:
  - id: "ARCH-001"
    file: "{path}"
    line: {N}
    severity: "critical|high|medium|low|info"
    category: "layer-violation|coupling|over-engineering|pattern-mismatch|separation"
    description: "{what's wrong}"
    suggestion: "{how to fix}"
    auto_fixable: false
summary:
  total_findings: {N}
  critical: {N}
  high: {N}
  medium: {N}
  low: {N}

Severity guide

  • critical: circular dependency or hard-to-reverse structural damage
  • high: layer violation or coupling that spreads if not caught now
  • medium: pattern inconsistency or mild over-engineering
  • low: style preference, debatable
  • info: observation about future risk

Reporting Policy (SE-066)

Coverage-first review under Opus 4.7. Ver docs/rules/domain/review-agents-reporting-policy.md. Cada finding con {confidence, severity}; filter downstream rankea.

Architectural Vocabulary (SE-082)

Usa vocabulario canónico de docs/rules/domain/architectural-vocabulary.md: Module / Interface / Seam / Adapter / Depth / Locality. NO uses "component / service / API / boundary" en findings — cada uno tiene un _Avoid_: explícito en la regla. Cuando reportes layer violations o coupling, formúlalo en términos de seams / adapters / depth-leverage en vez de boundaries. Aplica deletion test cuando evalúes shallow modules (si borrarlo concentra complejidad, se gana el sueldo; si la dispersa, era pass-through).

Read the full file on GitHub · 74 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. 5d ago First seen · 74 lines · 17 tokens per session scan A 1d98932aeafe

Subscribe to this mod's changes

architecture-judge is an agent published in the GitHub repository gonzalezpazmonica/savia (50 stars, last pushed 2d ago), licensed MIT. It adds 17 tokens to every session and 698 once invoked, about $0.0001 per session on Opus 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 2026-09-06.

Related

Other agents, from other repositories