An agent will finish a task, report success, and be completely wrong - not because it lied, but because nothing in its loop could ever have found out. Every law here is one way a system was made able to find out. All twenty-three were a real failure first.
# read it online, or take it with you git clone https://github.com/Aitherium/awknowledge.git # or offline, in your terminal, with the rest of the family pip install awkno && awkno law 5
Ten chapters to get from nothing to a working setup, then twenty-three laws. Filter by part, or search for the failure you are actually looking at. Every row opens the document.
The laws are grouped into seven parts. If you only have time for one, read Silence - those five cost the most per incident, because nothing tells you they happened.
Never used a coding agent? Ten chapters, about an hour, ending with one real verified change on your own repository. Covers local models and how to give an agent hands.
A standard nothing checks is a preference. How a preference becomes a property of the codebase - and why a checker nobody has watched fail is not a gate.
The expensive failures do not raise. They return 200, render correctly, log nothing, and leave the container healthy. A missing thing is indistinguishable from a thing nobody wanted.
A gate only works while people keep it switched on. Most gates die of being right too loudly - so narrow the rule, pin the count, ratchet down.
The gap between the code you wrote and the code that is running. A build can succeed and ship nothing; a live mount makes the file current and leaves the process stale.
Patterns no linter will catch, because they are semantic: the gate that fails open, the decision keyed on input the caller chose, the count that measures the wrong thing.
A setting is only real if it survives a restart and reaches the process that reads it. A cap in a file nothing loads is a wish; a decision several agents make together needs a channel a human can read.
Where the numbers stop meaning what they used to. A mixture-of-experts model has two budgets, not one, and a plan that adds capacity without asking which one it feeds adds nothing.
The codex is markdown. Clone it, read it, or point your agent at it - the laws are written to be loaded by an agent as easily as by a person.
git clone https://github.com/Aitherium/awknowledge.git cd awknowledge # the reading path ls path/ # the twenty-three laws ls laws/ # the same corpus, offline, searchable pip install awkno && awkno search silence
Not a reading list. The laws exist to be installed, and this is the loop that installs them. A codebase run this way gets harder to break over time without anyone having to be careful.
These projects are organs of the same system — the same identity plane, the same agent runtime, the same operational discipline. Follow any of them.
Every Aitherium repository publishes an aither-manifest.json beside its
page, and each surface reads all of them — so the network is browsable from any node in
it. A repository that publishes nothing shows as unknown, never as zero.