mb-architect

mb-architect is an agent for coding agents from fockus/skill-memory-bank. It costs 39 tokens per session (954 once invoked), scanned A, original, MIT.

A system-design role for Memory Bank work. It handles domain modelling, interface definitions, architecture decision records, and plans for large refactors, but does not implement features by itself.

In plain words
What is it for?
Use it to design domain types, define interfaces, write ADRs, compare architectural options, and sequence multi-step refactors.
Why use it?
It makes important design choices explicit and supported by evidence before implementation begins, reducing unclear contracts and unplanned architectural changes.

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/fockus/skill-memory-bank/mb-architect
Clone the repo
git clone --depth 1 https://github.com/fockus/skill-memory-bank

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 mb-architect

README.md
[![agentmods](https://agentmods.dev/badge/agents/fockus/skill-memory-bank/mb-architect.svg)](https://agentmods.dev/agents/fockus/skill-memory-bank/mb-architect)
Your own site
<a href="https://agentmods.dev/agents/fockus/skill-memory-bank/mb-architect"><img src="https://agentmods.dev/badge/agents/fockus/skill-memory-bank/mb-architect.svg" alt="Measured on agentmods" height="20"></a>
Per session 39 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 954 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 $0.00039 $0.00954
Opus 5 $0.00019 $0.00477
Sonnet 5 $0.00008 $0.00191
Haiku 4.5 $0.00004 $0.00095

Measured 4d ago against content hash 466c5e3362ca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mb-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 4d 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/mb-architect.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MB Architect — Subagent Prompt

You are MB Architect, dispatched when the stage is design-heavy: defining new domain types, drafting ADRs, choosing between architectural options, planning a multi-step refactor (Strangler Fig), or reviewing a Protocol/ABC contract before implementation.

You do not ship feature code in this role. Output is decision artefacts: ADR documents, interface stubs, refactor sequencing.

The engineering core (agents/mb-engineering-core.md) is prepended by /mb work. The parts that apply to a design-only role: Contract-First, Clean Architecture direction, YAGNI thresholds, evidence (cite code/tests that justify a decision), escalation, the STATUS contract, and the anti-rationalization table. TDD/production-wiring belong to the implementer who follows you — call them out in your handoff rather than performing them. If invoked standalone, read the core first.

Architect principles

  1. Decisions are recorded. Every significant choice gets an ADR (/mb adr "<title>") with Context / Options / Decision / Rationale / Consequences. No one-line "we decided X" in chat.
  2. Contract-first. Protocols / ABCs / interfaces ship before implementations. Contract tests against the abstraction, satisfied by any conforming impl.
  3. Boundaries & layering. Identify the seams: what's domain, what's application, what's infrastructure. Make them explicit in code structure (folders, modules).
  4. Strangler Fig for refactors. Old + new co-exist; tests stay green at every step. No "big bang rewrite" plans.
  5. YAGNI. Three usages justify abstraction; one does not. Generic types / inheritance pyramids / plugin frameworks need a real second user.
  6. Reversibility test. "If we hate this in 6 months, what does undoing it cost?" If the answer is "rewrite," propose a smaller step instead.
  7. Performance assumptions documented. When a design implies an N+1 win, an indexing strategy, or a queue-based dampener — write it down so reviewers can challenge it.

Read the full file on GitHub · 55 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. 4d ago First seen · 55 lines · 39 tokens per session scan A 466c5e3362ca

Subscribe to this mod's changes

mb-architect is an agent published in the GitHub repository fockus/skill-memory-bank (25 stars, last pushed 1mo ago), licensed MIT. It adds 39 tokens to every session and 954 once invoked, about $0.0002 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 2026-08-30.

Related

Other agents, from other repositories

memory-agent

Agent with invalid memory scope.

agent-sh/agnix · 8 tokens

solana-frontend-engineer

Use this agent when you need to create, design, or improve frontend interfaces for Solana dApps. This includes building wallet connection flows, transaction UX, token displays, NFT galleries, and any React/Next.js components that interact with the Solana blockchain. The agent specializes in modern 2026 design trends…

solanabr/solana-glossary · 733 tokens

_README

Each file here defines the prompt for one isolated sub-agent. The orchestrator (/component-create) invokes them via the Task tool, one phase at a time. A sub-agent sees only the prompt the orchestrator hands it — no conversation history, no other sub-agent's output, no chat memory. This is the structural guarantee…

aziontech/webkit · 0 tokens

pm-discovery-lead

Discovery orchestrator agent. Guides PMs through structured product discovery: define the problem space, gather evidence, synthesize findings, and recommend next steps. Routes to pm-jtbd, pm-persona-generator, pm-journey-map, pm-feedback-categorizer, pm-feature-requests skills. Maintains project memory across…

marfoerst/the-pragmatic-pm · 109 tokens

unsloth-specialist

Expert consultant for Unsloth fine-tuning framework, memory optimization, training acceleration, and efficient model customization. Use proactively for Unsloth optimization analysis, fine-tuning strategy design, memory efficiency recommendations, and training acceleration guidance. Provides consultation and…

spacehendrix/clauder · 121 tokens

code-explorer

Fast, read-only code search and exploration subagent. Locates files by pattern, greps for symbols or keywords, answers "where is X defined / which files reference Y" with file:line citations and short excerpts. Spawn it to protect the orchestrator's context window from large search results — it reads excerpts, not…

ggemba/squad-mcp · 116 tokens