cognitive-architecture-review

cognitive-architecture-review is a skill for Claude Code from impactbrussels/AINativeOS. It costs 160 tokens per session (1,438 once invoked), scanned A, original, Apache-2.0.

A method for checking whether an AI system depends on intelligence at its core or is mostly a conventional product with an AI feature attached. It scores the system's main parts and applies the Remove-the-AI test.

In plain words
What is it for?
Use it to audit an existing or partly designed AI architecture, find weak foundations, and decide what system loop should make the product improve over time.
Why use it?
It reveals fragile designs that work only in polished demonstrations and identifies missing feedback loops. Removing the AI shows whether the product's value truly depends on it.

Skill for Claude Code

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

Part of the ai-native-os plugin — 25 skills, 6 agents shipped together

Good fit Use it to audit an existing or partly designed AI architecture, find weak foundations, and decide what system loop should make the product improve over time.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/impactbrussels/ainativeos/cognitive-architecture-review
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.

Any agent
npx skills add impactbrussels/AINativeOS --skill cognitive-architecture-review
Clone the repo
git clone --depth 1 https://github.com/impactbrussels/AINativeOS

Made for: Claude Code.

Or install ai-native-os, the plugin that ships this one along with the rest of its 25 skills, 6 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 cognitive-architecture-review

README.md
[![agentmods](https://agentmods.dev/badge/skills/impactbrussels/ainativeos/cognitive-architecture-review/github.svg)](https://agentmods.dev/skills/impactbrussels/ainativeos/cognitive-architecture-review)
Your own site
<a href="https://agentmods.dev/skills/impactbrussels/ainativeos/cognitive-architecture-review"><img src="https://agentmods.dev/badge/skills/impactbrussels/ainativeos/cognitive-architecture-review/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 cognitive-architecture-review

Your own site · 80×15
<a href="https://agentmods.dev/skills/impactbrussels/ainativeos/cognitive-architecture-review"><img src="https://agentmods.dev/badge/skills/impactbrussels/ainativeos/cognitive-architecture-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 160 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,438 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 $0.00160 $0.01438
Opus 5 $0.00080 $0.00719
Sonnet 5 $0.00032 $0.00288
Haiku 4.5 $0.00016 $0.00144

Measured 10d ago against content hash 7e4a149d76a9, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

cognitive-architecture-review 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 10d 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.

skills/cognitive-architecture-review/SKILL.md · 99 lines

How it starts

The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Cognitive Architecture Review

A system that demos well tells you almost nothing. The demo runs on the happy path, with the founder steering, before any of the load lands. The question that decides whether you have a company is quieter: strip the AI out, and does anything of value remain? Most "AI products" survive the removal, which is the precise reason they get cloned and never command an AI-native price. This skill audits what you already built, names the pillar holding it up by luck, and finds the loop that should make it yours.

The method

Score the four pillars, find the missing loop, then run the Remove-the-AI test on the whole design. Full rubric, worked audit, and anti-patterns: references/review-rubric.md.

Step 1: Score each pillar against the rubric

Walk the system and grade each pillar 0 to 3 (absent, bolted-on, present, load-bearing). Use the red-flag column in the rubric to catch the pillar that is present on the slide and absent in the code.

Pillar The question that scores it Common red flag
Brain Which model reasons, and what is the cheaper fallback for bulk or failure? One frontier model on every call, no fallback, no eval
Memory What does the system know that the model does not? Memory is the context window; nothing persists between sessions
Planning How does it decide its moves, in order, before acting? One mega-prompt does retrieval, reasoning, and acting at once
Tools Which actions reach the world, and what is each blast radius? Tool calls with no guardrail on the irreversible ones

Step 2: Find the closed loop, or prove there is not one

Point to the single line where a customer's use this week makes the product measurably better next month. Trace it: use to capture to data to a sharper product. If you cannot draw that return arrow, record it as the headline finding. No loop means no flywheel means no moat, however good the pillars.

Read the full file on GitHub · 99 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 10d ago First seen · 99 lines · 160 tokens per session scan A 7e4a149d76a9

Subscribe to this mod's changes

cognitive-architecture-review is a skill published in the GitHub repository impactbrussels/AINativeOS (1 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 160 tokens to every session and 1,438 once invoked, about $0.0008 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.