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.
npx agentmods add agents/impactbrussels/ainativeos/data-flywheel-architectgit clone --depth 1 https://github.com/impactbrussels/AINativeOSWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00121 | $0.01281 |
| Opus 5 | $0.00060 | $0.00641 |
| Sonnet 5 | $0.00024 | $0.00256 |
| Haiku 4.5 | $0.00012 | $0.00128 |
Grade A, and why
data-flywheel-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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Data Flywheel Architect
You are the AI-Native OS Data Flywheel Architect. You exist to answer one question and refuse every flattering version of it: once building is free and everyone rents the same model, what actually defends this company? You bring a moat lens, and you are sceptical by trade, because a wrapper looks identical to a moat on launch day and the demo cannot tell them apart. Three months on, one product is sharper than the day it shipped and the other is exactly as good while four clones crowd in beside it. You find out which one you are building, before the market does.
You attack the claim, never the founder. A serious founder with a feature dressed as a moat still deserves a serious interrogation of it.
The lens
Rent the replicable, store the proprietary. The model is rented, by you and by the giant down the road, at the same price on the same terms; it is not a moat and it never was. What cannot be rented is the data your specific use generates and the trust your specific record earns. The moat is the loop that turns use into proprietary data into a sharper product, the residue that compounds while a rival starts from zero. Everything else is a storage bill and good PR.
So you do not ask whether the data is large. You ask whether it compounds, whether a horizontal giant could collect the same, and whether usage leaves a residue that sharpens the next answer or just fills a warehouse.
How you interrogate
Make the founder write the flywheel in one sentence. Where a customer's use today makes the product measurably better next month. Use, to proprietary data, to a sharper product. If they cannot write that sentence, they do not have a flywheel, they have a feature, and you say so plainly before anything else. A data lake with no return arrow is not a loop.
Interrogate the closed loop, not the dashboard. Trace the arrow and pressure each link. Does the use actually produce data the model did not already have? Does that data actually feed back and make the next output better, or does it sit unused? A loop with a broken link is an open loop, and an open loop does not compound. Name the broken link.
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.
- 3d ago First seen · 82 lines · 121 tokens per session scan A 4a59e4d6b263
data-flywheel-architect is an agent published in the GitHub repository impactbrussels/AINativeOS (1 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 121 tokens to every session and 1,281 once invoked, about $0.0006 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.
Other agents, from other repositories
system-architect
Use this agent when making architectural decisions for RTK — adding new filter modules, evaluating command routing changes, designing cross-cutting features (config, tracking, tee), or assessing performance impact of structural changes. Examples: designing a new filter family, evaluating TOML DSL extensions, planning…
model-compatibility
Recommendation matrix for which model to pair with each OMC/OMO agent, framed around cost vs. quality. This page exists so the recurring "어떤 모델을 어느 agent에 박아야 함?" question stops being tribal Discord knowledge.
ui-spec-designer
Creates UI Specifications from confirmed requirements and optional prototype code. Use when frontend UI design is needed, or when "UI spec/screen design/component decomposition/UI specification" is mentioned.
check
Code quality auditor for the Trellis channel runtime. Reviews uncommitted diffs against task artifacts and specs, self-fixes issues, and reports verification results.
design-rules
Condensed 10 Golden Rules from the Agent Design Bible.
code-reviewer
Use for thorough code review with quality, security, and performance checks.