Tracking: Valkey ecosystem tools listing (valkey-io/valkey-doc#475)

[kaya-abdullah]

Tracks getting LibreDB Studio named on Valkey's own ecosystem surface, so progress is visible from this repo (same pattern as #277 / #286).

Valkey connects through the Redis provider (ioredis), probed at Valkey 9.1.1 and recorded as Full in docs/providers/README.md.

Two tracks, and only one of them is a listing

1a. Data PR — cheap, but publishes nowhere today.

• Repo: valkey-io/valkey-doc
• File: tools/gui/github.com/libredb/libredb-studio.json (new)
• Precedent: tools/gui/github.com/uglide/RedisDesktopManager.json
• Schema: name, description, homepage, optional twitter
• CI: DCO (probot), Spellcheck, check-links, man-page build. No CLA — commits signed off with -s.

The catch, verified 2026-08-26: nothing renders this data. valkey.io/tools/ is 404, the sitemap has no page generated from the tools/ tree, the site repo builds only topics and clients, and no script inside valkey-doc itself reads tools/ either — not the Makefile, not utils/. The tree is inherited from redis-doc. So this is queue positioning, not a traffic channel, and there is no URL to add to the tracker.

Two claims in our own plan were wrong and are corrected here: check-links does not validate the homepage (it globs {topics,commands}/* only and skips absolute URLs), and tool_types.json sits at the repo root rather than under tools/.

1b. Getting a page to exist — the actual prize.

• valkey-io/valkey-doc#405 "Proposal: Add a Tools & Ecosystem page to valkey.io" is still open, and its author offered to write the first PR. The first category proposed is "GUI & management tools".
• Pattern to copy: the AI Libraries page — data PR valkey-doc#459, template PR valkey-io.github.io#602, site issue #595. So it is a double PR: data to valkey-doc, Zola template to the site repo.
• Maintainer conditions (stockholmux, Add a Hindi README (README_hi.md), covered by the localized-README drift guard #595): one broad page rather than a narrowing name, data-driven not static, a date field per entry to expose staleness, an issue template plus an inclusion standard.
• This is page-building, not a listing submission — over a week of work. For comparison, the ClickHouse GUI page sends 20 referrals a day, and this channel is at zero.

Status

valkey-io/valkey-doc#475 is open. DCO passes. Spellcheck, check-links and the man-page build are all sitting at action_required: valkey-doc gates workflows for first-time contributors, so a maintainer has to approve the run. Nothing to do on our side, and none of the three touches this file anyway.

The PR body states the one caveat outright rather than leaving it to be found: the overview reads redis_version from INFO, so a Valkey 9.1.1 server displays 7.2.4.

Checklist

• Target file, schema and CI requirements verified against the live repo
• Publishing surface checked before submitting (it failed the check — see above)
• docs/providers/redis.md names Valkey (docs(providers): the Redis doc names its three relatives, Valkey first #505)
• Data PR opened against valkey-doc: Add LibreDB Studio to the GUI tools list valkey-io/valkey-doc#475 (opened 2026-08-27)
• Data PR merged
• Decide whether to take on 1b

[kaya-abdullah]

Update 2026-08-28

valkey-doc#475 is approved but not merged. stockholmux left an LGTM on 27 Aug, then commented that
this part of the docs is not currently rendered and asked whether we would rather write a blog post.
Answered: merge it whenever it suits, and yes to the post. That confirms track 1a on its own is queue
positioning, exactly as this card recorded.

The blog became the real channel, and both steps are submitted.

• Proposal issue: [BLOG POST] Browsing a Valkey keyspace safely: SCAN, INFO, and a read-only ACL user valkey-io/valkey-io.github.io#648, with the blog-post label.
  An earlier attempt, docs: four sites credit SQLBaseProvider with placeholder generation it does not have #647, was opened from the command line and the API silently dropped the label
  because we have no push access on that repo. It was closed with a note. Open issues there through
  the web template, not gh, or the template labels never apply.
• Post PR: Add blog post: Browsing a Valkey keyspace safely valkey-io/valkey-io.github.io#649. Two new files, content/blog/2026-08-28-browsing-a-valkey-keyspace-safely.md
  and content/authors/kaya-abdullah.md. DCO passed, @valkey-io/content-creators assigned by CODEOWNERS,
  mergeable, CodeRabbit still running at the time of writing.

The post is not a product pitch, by their rules. CONTRIBUTING-BLOG-POST.md in the site repo is
binding: 500 to 1200 words, one sentence per line in the source, language tags on code fences,
acronyms expanded on first use, a closing call to action, and no promoting a product. The install
commands were cut for that last rule. The angle is SCAN instead of KEYS and an ACL user that bounds
the session, with the multi engine point kept to one sentence. Valkey Admin, the project's own GUI
announced 12 May 2026, is named in the post so it does not read as competing with it.

Everything in the post was measured, not asserted. Valkey 9.1.1 from valkey/valkey, default
configuration: the ACL grant, the panels answering as the restricted user, and the NOPERM refusal on
a write. The site was built locally with zola 0.22.0, the version the deploy workflow pins.

Local clone: ~/projects/valkey-io.github.io, branch blog/browsing-a-valkey-keyspace-safely, commit c4b18cb.

A product bug fell out of this and is not fixed yet. The Redis provider reads redis_version from
INFO, but Valkey publishes server_name:valkey and valkey_version:9.1.1 on the next two lines, so a
Valkey server shows as 7.2.4. Measured the same day: DragonflyDB publishes dragonfly_version and
Garnet publishes garnet_version, so three of the four relatives are misreported, while our own docs
record this as a Garnet-only problem in docs/providers/redis.md, docs/providers/README.md and
src/lib/db/compatibility.ts. Only KeyDB genuinely publishes nothing. Handed to cevheri. Note for
whoever writes the fix: Redis, KeyDB and Valkey all publish gcc_version, so a rule that takes any
*_version field other than redis_version reports the compiler on plain Redis.

Waiting on three things: the #475 merge, the content-creators review on #649, and the CodeRabbit pass.