Review and review issues
How store review talks to you. Each problem is an issue with its own thread; answer it on /dev/submit, with 0x issues, or with the API.
Every store submission is reviewed by 0x before it goes live. When a reviewer finds a problem,
they raise a review issue: one problem, with its own thread between the reviewers and you.
You answer where you already work: the submit page, the 0x CLI or the API. The reviewer then
resolves the issue, marks it won't-fix, or reopens it.
#An issue
| field | meaning |
|---|---|
category | security, source-mismatch, content, metadata, technical, other, or rejection (the reason a version was rejected) |
severity | info, needs-change or blocking. Only open blocking issues stop an approval. |
title and messages | What the reviewer wrote. Titles are one line of up to 200 characters; messages are up to 4000. |
refs | Where in your source: up to 20 of {file, line?, rule?} |
origin | What raised it: scan, ai, file, rejection or manual, and the rule where there is one |
state | open → replied → resolved / wontfix; a closed issue can be reopened |
fixed_in_version | The version you said fixes it |
How the state moves:
| action | from | to |
|---|---|---|
| you reply | open, replied, reopened | replied |
| reviewer replies to you | replied | open (your turn again) |
| reviewer resolves / marks won't fix | open, replied, reopened | resolved / wontfix |
| reviewer reopens | resolved, wontfix | reopened |
A closed issue takes no replies (409). Only a reviewer can reopen it. Reviewers appear as "0x review". You never see internal notes, reviewers' names
or the AI review's raw findings: an AI finding reaches you only as an issue a reviewer raises,
in their words.
A rejection keeps its reason as the version's review note, and also opens a rejection issue
you can answer.
#Answering on /dev/submit
https://api.0x.limo/dev/submit shows Review issues first, with a red "N new" badge (and
"(N)" in the tab title). Each issue opens to its thread, with a reply box and a Fixed in
choice (your listing's later versions). Link to an issue with /dev/submit#issue-<id>. The
badge refreshes every 2 minutes while the tab is visible.
#Answering with the CLI
0x issues # every issue on your listings, active first; [new] marks unread
0x issues my-game # one listing's issues
0x issues show 42 # one issue and its thread (marks it read)
0x issues reply 42 "Removed the eval in loader.js; fixed in version 4" --fixed-in 4
0x issues --json # machine-readable
0x submit status # your submissions, with issue countsThe CLI strips control characters and bidi overrides before printing anything a reviewer wrote.
#Answering with the API
Use an access token, or a publish token from 0x login (publish tokens reach /v1/hosting/* and
/v1/listings/* only).
curl -s -H "Authorization: Bearer $ZEROX_TOKEN" https://api.0x.limo/v1/listings/issues
curl -s -H "Authorization: Bearer $ZEROX_TOKEN" https://api.0x.limo/v1/listings/my-game/issues
curl -s -H "Authorization: Bearer $ZEROX_TOKEN" https://api.0x.limo/v1/listings/issues/42
curl -s -X POST -H "Authorization: Bearer $ZEROX_TOKEN" -H 'content-type: application/json' \
-d '{"message": "Fixed: the loader no longer uses eval.", "fixed_in_version": 4}' \
https://api.0x.limo/v1/listings/issues/42/reply| call | returns |
|---|---|
GET /v1/listings/issues | {items, open, unread, awaiting_you}, active issues first (up to 200) |
GET /v1/listings/{id}/issues | One listing's issues (numeric id or slug) |
GET /v1/listings/issues/{issue} | The issue and its thread. Marks it read. |
POST /v1/listings/issues/{issue}/reply | {message, fixed_in_version?} |
GET /v1/listings | Each listing carries issues: {open, blocking, unread, awaiting_you} |
Another developer's issue answers 404. Limits: one reply every 20 seconds per account, with a
burst of 10; at most 200 messages per issue and 100 active issues per listing.
#Notifications
There is no email or webhook yet. You're notified in-app: the badge on /dev/submit, the issues
counts in GET /v1/listings, and [new] in 0x issues.
#Tips
- Fix the problem in a new version, resubmit it, then reply with Fixed in that version. That is the fastest path to approval.
- Issues raised from a source scan quote the file and line. If the finding is a false positive (for example vendored upstream code), say so and point at it.
- Keep upstream code under
vendor/orthird_party/, so the review reads your own code first (see Source integrity).
