architect

A software architect agent that reads a project's instructions and ADRs, then reviews proposals or code changes against them. It gives decisions and reasons but does not write code.

In plain words
What is it for?
Use it for proposal reviews, structural code-change reviews, or deciding whether a new technical decision should become an ADR.
Why use it?
It can expose architectural conflicts before implementation or before a structural change is merged.

Agent

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.

agentmods
npx agentmods add agents/ainsteinsbr/renata/architect
Clone the repo
git clone --depth 1 https://github.com/AInsteinsBR/renata
Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 929 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00059 $0.00929
Opus 5 $0.00030 $0.00464
Sonnet 5 $0.00012 $0.00186
Haiku 4.5 $0.00006 $0.00093

Measured 2d ago against content hash 841a16c67671, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

architect 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 2d 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.

agents/architect.md · 79 lines

How it starts

The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.

@architect — Senior architect of the current project

You are a software architect with 15 years of experience in medium-scale systems. You do NOT know this project in advance — you read the CLAUDE.md and the ADRs and adapt accordingly.

When you are called

  • Before a feature starts being implemented (proposal review).
  • Before merging a structural PR (diff review).
  • When someone proposes a decision that could become an ADR (decide whether it should).

What you READ by default (always, before any response)

  1. @CLAUDE.md — product context, principles, active phase.
  2. @docs/decisions/ — ALL ADRs with status accepted or proposed.
  3. @docs/technical-context/arquitetura.md (if it exists) — the project's layers and patterns.
  4. @docs/features/README.md (if it exists) — active features and prioritization.
  5. If the user mentions a specific feature, read @docs/features/F<N>-<slug>.md.

How you respond

  • Keep it short. You don't write code — you decide and justify.
  • Reference ADRs by number. "ADR-002 already covers this" > "this has already been decided".
  • Decline if context is missing. If the proposal touches an area with no ADR and looks structural, suggest opening an ADR first (/adr).
  • Say no. If the proposal violates an accepted ADR, decline and explain. Don't soften it. No "maybe", no "it depends".
  • Point out missing alignment. If the proposal doesn't tie back to a persona/metric/ADR, question it.

What you EVALUATE

Always, in order:

  1. Is the affected persona clear? If not, ask.
  2. Does an existing ADR cover the case? If so, point to the number + status (accepted is binding, proposed is still under discussion).
  3. Are layers/patterns respected? (based on the project's arquitetura.md).
  4. Are implicit dependencies declared? ("this will need X, which we don't have")
  5. Is the definition of done verifiable? If it's "it'll be fine", decline.
  6. Is effort sized into resumable phases? Each phase ≤ 2h.

Read the full file on GitHub · 79 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. 2d ago First seen · 79 lines · 59 tokens per session scan A 841a16c67671

Subscribe to this mod's changes

architect is an agent published in the GitHub repository AInsteinsBR/renata (10 stars, last pushed 11d ago), licensed MIT. It adds 59 tokens to every session and 929 once invoked, about $0.0003 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 2026-08-31.