Trust score
Every Wakehood profile shows a trust score. It is a single number, and the most useful thing this page can do is tell you precisely what it is made of and precisely what it cannot tell you.
The formula
trust score = (10 x published agents) + (1 x whole days since account creation)
That is the entire calculation. There are no hidden terms, no weights, no decay, and no discretionary adjustments. The number shown on a profile is always reproducible from two public facts: how many agents that key has published, and when it first appeared.
Details that matter:
- Days are whole days, floored. An account created 23 hours ago contributes 0 days.
- The score never goes negative.
- The UI always shows the breakdown, so the number can be explained rather than just asserted.
A worked example
A developer creates an identity, and 30 days later has published 5 agents.
| Component | Calculation | Points |
|---|---|---|
| Published agents | 5 agents x 10 | 50 |
| Account age | 30 days x 1 | 30 |
| Total | 80 |
Now compare two profiles that both score 80:
- One published 8 agents on day zero and is 0 days old.
- One published nothing and has existed for 80 days.
Both score 80. Neither number says anything about whether their code works. That is not a flaw in the example. That is the point of the next section.
What it measures
Participation and longevity. It answers, roughly: has this key shown up, and has it been around for a while?
That is a genuinely useful signal against one specific thing, which is disposable identities. Spinning up a fresh keypair costs nothing, so a brand new account with a score of 0 tells you that you are looking at something with no history. A key that has published steadily over months has at least accumulated a public track record of what it published, which you can go and read.
What it does not measure
Read this list carefully, because a number next to a name invites people to read meaning into it that is not there.
It is not a measure of trading skill. Wakehood collects no performance data in Phase 1. No returns, no drawdowns, no backtests, no live results. The score cannot reflect performance because performance is not an input.
It is not a measure of code quality or safety. Nobody at Wakehood reads, audits, tests, or runs the code. The score goes up when a file is uploaded, not when the code is good.
A high score does not mean an agent is good. It does not even mean the agent works. A profile with a score of 500 might have published 50 broken scripts, or 50 identical ones. The score counts the act of publishing. It does not judge it.
It is not a warranty, a rating, or an endorsement. It is not investment advice, and it is not a reason to run someone's code or to give it access to funds.
The only thing that tells you whether an agent is worth anything is reading the agent. Wakehood never runs your code, so a listing carries no execution history and no verification. The code is on IPFS and the CID is on the detail page precisely so that you can go and read it yourself.
Why it is deliberately this dumb
The obvious objection is that a score which ignores performance is a nearly useless score. That is close to true, and it is still the right call for Phase 1. Here is the reasoning.
The alternative is a score derived from performance claims, and in Phase 1 there is no way to verify a performance claim. Wakehood does not execute agents, so it observes no trades and no results. Any performance number would therefore be a number the author typed in.
A score built on self-reported, unverifiable returns is not a neutral score. It is actively harmful, because it launders a claim into an apparently objective metric. "247% APY, trust score 94" reads as though a system checked something. Nothing was checked. The number would make the platform complicit in a claim it cannot stand behind, and the people harmed by that are exactly the people the score is supposed to protect.
So Phase 1 ships a score that is honest about being shallow: two inputs, both objectively true, both publicly checkable, and a name that is qualified everywhere it appears. A dumb score you can fully see through is safer than a sophisticated one you cannot.
Where trust is going
On the roadmap, and not built:
- Phase 3 intends to derive trust from real trading performance, alongside community staking, so that reputation is backed by something with a cost attached rather than by the act of uploading a file.
- Phase 5 proposes staking
$WAKEon an agent, which would let developers and the community put capital behind their conviction, and an insurance pool that could be slashed when trust is betrayed.
Neither exists today. Until they do, the number on a profile means exactly what the formula at the top of this page says it means, and nothing more.