//00 agentic ai lab · portugal
AI agents in production,
built, observed and attacked
with evidence.
AI agents in production: how they are built, operated, evaluated, governed and attacked, tested on a real stack. Flagship area: AI Ops Sec, agent security and observability.
Lab areas
7AI Ops Sec
Agent security and observability
Agents and orchestration
How to build an agent that survives production
Automation and ROI
Business processes handed to agents, with the numbers
Evaluation and quality
Knowing whether the agent got it right before the client does
Governance and regulation
What the law already asks of agents, read by an operator
Models and infrastructure
Local and commercial models, and the infrastructure that serves them
Tokenization
The token as the economic unit of AI: what it costs, what it saves, what it hides
Latest articles
06- AI Ops Sec
J-Lens: reading what the agent thinks before it acts, and what that changes in observability
Anthropic showed an internal space where the model keeps what it has in mind. I ran the lens on four synthetic prompts against a 27B Qwen and designed the third lane of agent observability.
- AI Ops Sec
The danger is not the connector. It is what gets into the model, and what it can do next
'Is this MCP secure?' has an answer. The missing question is another: what did I just authorise into a model that can also act? Two attack chains, what is at stake, and observability over MCPs.
- Tokenization
The five thousand dollars you never see on the invoice: why cache is your hidden balance sheet
Your invoice shows what you spent. It never shows what you avoided spending. In my own operation, what I avoided was four and a half times what I paid, and it appears nowhere.
- Tokenization
The agentic multiplier: where your tokens REALLY go
In an agentic system, most of the tokens you pay for are never seen by anyone. It is called alpha, and in my own operation it reaches 11.1×. Here is how to measure it before the bill measures it for…
- Tokenization
The 35% you don't see: why "price unchanged" can be a disguised increase
The vendor announces "price unchanged". Your bill goes up 35%. How the tokenizer changes between versions and yields more tokens for the same text, and how to measure the factor in your operation.
- Tokenization
The 2026 Token Crunch: Why Your AI Bill Exploded (and How to Cut It 70–90%)
Per-token price fell from $10 to $2.50/M in a year. Bills multiplied because agents burn 5–30× more tokens. Five documented levers cut 70–90% of cost.
How this site is built
- 01Code and articles live in a public repository. Every change lands through a reviewed pull request.
- 02No trackers, no cookies, no third-party scripts. Static HTML served from a CDN.
- 03Every article lists its sources and what changed after publication.
- 04Mistakes are fixed in public, in the history of the article itself.