GloriousFlywheel

GloriousFlywheel Downstream Adoption Inventory Matrix 2026-04-16

Snapshot date: 2026-04-16

Purpose

Turn the downstream-adoption lane into an explicit inventory instead of a vague “some repos use the runners” story.

GitHub owner: #210

Current Direct Usage Reality

GitHub code search on 2026-04-16 shows:

32 workflow-file hits across 6 repos in tinyland-inc
19 workflow-file hits across 6 repos in Jesssullivan
combined:
- 51 workflow-file hits
- 12 repos using runs-on: tinyland-... labels

This is the direct label-usage surface. It is larger than the original 5-repo first-tranche assumption.

It also understates the real adoption surface.

Literal label search misses repos whose self-hosted policy is expressed through repository variables rather than hard-coded runs-on: tinyland-* strings. Jesssullivan/MassageIthaca is the clearest current example: its core CI already uses repo-owned honey runner labels via PRIMARY_LINUX_RUNNER_LABELS_JSON, even though its remaining beta escape hatches were still pinned to ubuntu-latest.

Important Distinction

There are now three separate scopes:

direct label-usage surface: 12 repos
named variable-driven dogfood surface: repos like Jesssullivan/MassageIthaca that may not show up in literal label code search but still consume repo-owned GloriousFlywheel runners
operational first migration tranche: 5 repos
default owned-repo policy footprint: 264 non-fork repos across tinyland-inc and Jesssullivan

Near-term rollout should still start with the first execution tranche plus the named dogfood blockers, but the policy boundary is now the full owned-repo surface rather than only the literal label hits.

Direct Inventory

Repo	Owner class	Workflow-file hits	Current status	Direct action-path evidence	Suggested tranche	Migration shape
`tinyland-inc/GloriousFlywheel`	platform source	4	self-dogfood and release surface	self-reference visible in release workflow	tranche 1	keep as truth source, remove drift first
`tinyland-inc/blahaj`	infra consumer	12	strongest adjacent operational dependency	none visible in code search	tranche 1	stabilize label and cache contract
`tinyland-inc/lab`	operator tooling consumer	10	heavy validation and cache-adjacent consumer	none visible in code search	tranche 1	clarify runner vs cache vs MCP dependency
`tinyland-inc/tinyland.dev`	product repo consumer	4	representative GitHub Actions consumer	none visible in code search	tranche 1	publish one canonical shared-runner pattern
`Jesssullivan/XoxdWM`	user canary	8	strongest named user-repo consumer	current `tinyland-inc/GloriousFlywheel/...` path visible	tranche 1	validate public contract and action-path migration
`tinyland-inc/betterkvm`	org product repo	1	light direct runner-label consumer	none visible in code search	tranche 2	classify workload and shared-label need
`tinyland-inc/ironclaw`	org product repo	1	light direct runner-label consumer	none visible in code search	tranche 2	classify workload and shared-label need
`Jesssullivan/fuzzy-crush`	user repo	2	light direct runner-label consumer	none visible in code search	tranche 2	decide whether it is a supported canary
`Jesssullivan/remote-juggler`	user repo	3	medium direct runner-label consumer	none visible in code search	tranche 2	classify as container/build or infra consumer
`Jesssullivan/tinyclaw`	user repo	3	medium direct runner-label consumer	none visible in code search	tranche 2	classify as product-repo or canary consumer
`Jesssullivan/tinyland-hexstrunk`	user repo	1	light direct runner-label consumer	none visible in code search	tranche 2	classify shared-label need and de-risk
`Jesssullivan/yt-text`	user repo	2	live shared-runner consumer; healthy pickup on `tinyland-nix`	none visible in code search	tranche 2	optimize workflow shape, bootstrap Nix explicitly, stop using `tinyland-nix` for Pages deploy, and use it as the first `tinyland-nix-heavy` Rust/Clippy dogfood canary; see yt-text dogfood playbook
`Jesssullivan/MassageIthaca`	dogfood product repo	variable-driven	repo-owned `honey` lane active; beta validation was still escaping to `ubuntu-latest`	runner policy currently lives in `PRIMARY_LINUX_RUNNER_LABELS_JSON`, not literal label strings	named dogfood blocker	move beta and alpha post-deploy validation onto the repo-owned runner lane and use it as the proof that product repos do not need hosted-runner escape hatches

First Tranche Contract

The first migration tranche should still stay at 5 repos:

GloriousFlywheel
blahaj
lab
tinyland.dev
XoxdWM

Reason:

these repos cover the source-of-truth platform, the cluster-adjacent infra consumer, the operator-tooling consumer, one representative org product repo, and one user-owned canary
they are enough to validate docs, labels, cache behavior, and action-path migration without widening the blast radius too early

Second Tranche Candidates

The other 7 directly exposed repos should be treated as named follow-on candidates, not ignored:

betterkvm
ironclaw
fuzzy-crush
remote-juggler
tinyclaw
tinyland-hexstrunk
yt-text

These are important because they prove the direct runner-label surface is already broader than the platform and infra repos alone.

Action-Path Reality

Current code-search evidence from this pass:

current tinyland-inc/GloriousFlywheel/... workflow/action path usage is visibly downstream in Jesssullivan/XoxdWM
no current visible Jesssullivan/GloriousFlywheel/... workflow references were found in org or user workflow code search

Meaning:

the strongest downstream action-path migration target right now is XoxdWM
the larger direct adoption surface is currently driven more by shared runner labels than by direct composite-action references

Immediate Outputs Needed Under `#210`

a repo-by-repo classification table with:
- label family used
- cache expectations
- secret/bootstrap shape
- rollback path
one migration checklist for tranche 1
one decision note for which tranche-2 repos become named support canaries
one public downstream contract page once tranche 1 is stable