Moderation outcomes
What happens to a submission
A queue that only shows pending items tells you nothing about the standard. Below is every gate outcome in the 10 sample submissions that ship with this build, aggregated and explained, plus the reviews you have made yourself — read back from your own browser.
Honest limit: this build has no server, so there is no live public ledger of real submissions. The aggregate below is computed from the sample queue, which is stable and can be audited against the admin queue in the same build. Your decisions stay on this device and never enter these numbers.
10
sample submissions
2 element · 3 section · 2 animated · 3 prompt
4
every gate clean
40% of the queue
1
blocked by safety
rejected by default, no human override
90.9
average score
a11y audit for components, fidelity for prompts
Lint gate
Style and structure rules — the failures are fixable, not disqualifying.
Safety gate
Sandbox behaviour for components, model-run integrity for prompts. A failure blocks the entry.
The three decision paths
4
Clean — publishable
Both gates pass. The reviewer confirms the remix note matches what actually changed, then it publishes under the author's name.
5
Needs review — a conversation
A warning is not a rejection. The reviewer either asks for a fix or approves with the warning recorded on the page.
1
Blocked — rejected by default
A safety failure does not go to a human first. The submitter gets the reason and can resubmit; nothing publishes under their name meanwhile.
Entry by entry
SaaS pricing — glassmorphism landing
SUB-1026 · prompt · @pixelparlor · Next.js · 3 models
Soft ripple text field
SUB-1025 · element · @lena.dev · React
Lava lamp blob hero
SUB-1024 · section · @noir.studio · HTML/CSS
Indie magazine cover hero prompt
SUB-1023 · prompt · @glyph.rgb · HTML · 3 models
Coin flip loader
SUB-1022 · element · @karina_ui · React
Scroll-linked hue nav
SUB-1021 · animated · @tttyping · Vue
Glass stat card trio
SUB-1020 · section · @pixelparlor · HTML/CSS
Aurora pricing toggle
SUB-1019 · animated · @unknown_usr · React
Beauty routine-builder prompt
SUB-1018 · prompt · @studio.ceres · HTML · 3 models
Checkout stepper
SUB-1017 · section · @monoflow · React
9 contributor handles appear across these records — the sample roster used everywhere else on this site. With real submissions, this table is exactly what a public decision log would look like, and the counts above would be the aggregate.
Your reviews on this device
Read from the same store the admin queue writes to. Nothing here is uploaded, so nothing here can appear in a public statistic.
Loading your reviews…
Decisions recorded —. Your accept rate is yours alone; the aggregate on this page is computed from the sample queue that ships with the build, so it cannot be skewed by anything a visitor does.
What we will not publish
- Rejection rates for named people. An outcome belongs to an entry, not to a reputation.
- A “submissions this month” counter that only ever grows — a number nobody can verify is decoration.
- Appeal traffic, reviewer names or turnaround time promises we cannot keep at this size.
What would make this page real
- A submissions table with the gates run server-side, so the aggregate is computed, not transcribed.
- Decisions written as append-only rows — corrections become new rows, never edits.
- The same public view for authors: your own entry history, with the gate output attached.
Until then this page shows the sample queue honestly rather than an empty ledger pretending to be one.