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 skills add impactbrussels/AINativeOS --skill design-the-mvagit clone --depth 1 https://github.com/impactbrussels/AINativeOSWrote 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.
[](https://agentmods.dev/skills/impactbrussels/ainativeos/design-the-mva)<a href="https://agentmods.dev/skills/impactbrussels/ainativeos/design-the-mva"><img src="https://agentmods.dev/badge/skills/impactbrussels/ainativeos/design-the-mva/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.
<a href="https://agentmods.dev/skills/impactbrussels/ainativeos/design-the-mva"><img src="https://agentmods.dev/badge/skills/impactbrussels/ainativeos/design-the-mva.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00154 | $0.01289 |
| Opus 5 | $0.00077 | $0.00645 |
| Sonnet 5 | $0.00031 | $0.00258 |
| Haiku 4.5 | $0.00015 | $0.00129 |
Grade A, and why
design-the-mva 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 9d 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Design the MVA
The cheapest thing in your company is now a feature. You describe one in a sentence and it works by lunch, which is exactly the trap dressed as a gift. So founders pile up panels and integrations and call it a product, while the one question that decides whether the company lives goes untested: will a real person let a machine own the job the whole thing depends on. Ship the smallest agent that proves that loop, not the feature pile that flatters you.
The method
Full method, the Goldilocks scoring filters, a worked MVA scope, anti-patterns, and a copyable
template: references/mva-method.md.
Step 1 - Find the Goldilocks task
Pick one task that is high-volume, guideline-based, and low-risk, all three at once. High-volume, so the loop runs often enough to save real time and throw off real signal. Guideline-based, so there is a knowable right answer a human can check at a glance. Low-risk, so a wrong answer is an annoyance and not a harm. Your highest-value task is usually your highest-risk one. Write that on the roadmap for later. Earn trust on the boring task next to it first.
Step 2 - Define the Minimum Viable Agent
The Minimum Viable Agent is the smallest agent that proves the intelligence loop closes: task in, agent acts, outcome out, human accepts or corrects, the system records the choice. The deliverable is not a screen. It is a working loop around one task, and it can be thin on the surface: a text box, an email reply, a row someone approves. The intelligence is the product.
Step 3 - Name the one high-risk hypothesis it tests
State the trust bet so it can come back no: users in role R, facing task T, will accept the agent's output without redoing it at least M percent of the time. One hypothesis, not a multi-feature dashboard. If a proposed build does not map to a bet that can fail, it is scope creep, however reasonable it sounds.
Step 4 - Add the self-healing fallback
Decide in one bounded sentence what the agent does when it is unsure or the model fails: retry a set number of times, switch to a backup, then drop to a fixed non-AI path and escalate to a named human. A pilot user who hits a dead end twice goes back to doing it by hand. You lose them on reliability, not capability.
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.
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.
- 9d ago First seen · 69 lines · 0 tokens per session scan A a837ec0ffe58
design-the-mva is a skill published in the GitHub repository impactbrussels/AINativeOS (1 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 154 tokens to every session and 1,289 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.
Other skills, from other repositories
remember
Review reusable project knowledge and decide what belongs in project memory, notepad, or durable docs.
docx-comment-reply
Reply to comments (批注) in Word .docx/.doc files: extract comment context, draft replies, write threaded replies back, and validate OOXML.
explaining-machine-learning-models
Explain trained machine learning models through feature attribution, local explanations, and behavior summaries. Use as an explicit/manual helper once a model already exists, not for training ownership, leakage auditing, or general ML strategy selection.
config-validator
Validate AIWG configuration files and project setup for correctness and completeness.
aiwg-regenerate-copilot
Regenerate copilot-instructions.md for GitHub Copilot with vendor-specific content only.
checkpoint
Create, list, or recover mid-workflow checkpoints so interrupted work resumes from a known-good position.