Skip to content
SHC Docs

App capabilities

Layer 6. Services. Services

Two axes, and they are not the same axis.

  • Upstream capability: what the APPLICATION can do. Declared in each app’s features: block, by whoever wrote the app’s features: block. Changes only when upstream changes.
  • SHC integration: what the app wires up. Derived, on every run of this generator, from the files on disk. Never declared, because a declared answer to a question the filesystem already answers is an answer that rots.

★ A gap is never a mark against the app. An app is scored on what the APP supports; where SHC has not caught up, that is SHC’s debt and this page is the bill. Nothing here is a score: no weights, no totals. The scoring lane consumes -format json.

★ unknown is a value, not an absence, and it must never be read as absent. Most of this catalogue is undeclared today.

Count
Apps108
Apps with a features: block69
Apps with none (every value unknown)39
Feature keys in the vocabulary16
…of which SHC has an integration surface for9
…of which SHC has no surface at all7

★ That last row is the headline finding of this page. For 7 of the 16 capabilities a shopper cares about, SHC cannot wire anything even in principle: there is no plug, no socket and no derivation anywhere in the tree for ldap, mfa, scim, multiuser, multi_tenancy, audit_log, mobile_app. Those values can only ever be declared, which is the strongest argument that the declared axis is not redundant with the derived one.

The vocabulary, and why each entry earned its place

Section titled “The vocabulary, and why each entry earned its place”

features: is a closed vocabulary. A name that is not on this list is completely ignored: not rendered, not scored, not an error. That is what separates a feature from a tag: a tag is whatever the developer typed, a feature is a fact a shopper can filter on.

The owner named it first, and it has the catalogue’s widest spread: 41 apps ship an oidc.plug.yaml and 3 a saml.plug.yaml, while every one of the 6 apps/enterprise/ apps is in that bucket BECAUSE its SSO costs money. jellyfin, uptime-kuma and agent-canvas have none at any price. Nothing else in this list separates apps as hard.

SHC surface: an oidc.plug.yaml or saml.plug.yaml in the app. Failing that, a published oidc.socket.yaml (the app IS the IdP). Failing both, access: gated: SHC’s forward-auth gate, which puts Keycloak in front of an app that has no login of its own (a SUBSTITUTE, not a wiring).

The owner’s named fallback: “if it doesn’t support single sign on, does it at least support some kind of identity manage thing like LDAP?”. It is ALSO the cleanest proof that the two axes are different: the catalogue contains ZERO ldap plugs and zero ldap sockets, so this value can only ever be declared, never derived. gitlab, mailcow, penpot, metabase and midpoint all speak LDAP upstream and SHC wires it for none of them.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

Varies and matters: vaultwarden, sonarqube, fleet, midpoint, hyperswitch and librechat carry their own TOTP/WebAuthn, while nextcloud ships it only as the twofactor_totp and otpmanager ADD-ONS, which is the evidence that it is not core there. The operator’s real question is what happens when the app is NOT behind the gate, because then the app’s own MFA is the only MFA.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

The single most common thing left behind a paywall when SSO itself is free: fleet (Premium), ciso-assistant (Pro), sonarqube (Server tier) and windmill (EE) all give away login and charge for automatic joiner/leaver provisioning, per the licensing audit. It is the capability an estate with staff turnover actually needs, and exactly one file in the whole tree mentions it today, so nearly every value will be unknown. That is the tri-state doing its job rather than a reason to drop the key.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

Distinct from multiuser: uptime-kuma has one shared admin, gitea has org/team permissions, and DefectDojo 3.x moved RBAC itself behind Pro. An operator handing an app to a department needs to know which of the three they are buying.

SHC surface: the app’s SSO plug names a group or role claim, so SHC maps IdP groups onto the app’s roles rather than merely logging people in.

The owner’s flagship bar (“unlimited users”) made per-app and per-fact. Four apps fail it for three different reasons (uptime-kuma has no user model at all, agent-canvas has one shared session key, affine caps a workspace at 10 members and windmill at 50 accounts), and a shopper cannot tell any of that from the tile.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

SHC is itself multi-tenant, so “can one instance serve two isolated organisations” decides whether an estate needs one deployment or twenty. It varies AND it interacts with the licence: dify’s modified-Apache terms forbid multi-tenant operation without written authorisation, and windmill’s free tier caps workspaces at 3.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

The owner named email integration. 39 apps ship an smtp.plug.yaml and 51 do not, and for a self-hoster “does it need me to stand up a mail path” is a deployment prerequisite, not a nicety.

SHC surface: an smtp.plug.yaml in the app, or a published smtp.socket.yaml.

Split from email_outbound because it is a different capability with a different blast radius, and the catalogue proves the split: exactly 3 apps ship an imap.plug.yaml (frappe, mautic, nextcloud) against 39 for smtp. Ticketing, CRM and document-intake apps live or die on it; everything else never asks.

SHC surface: an imap.plug.yaml in the app, or a published imap.socket.yaml.

The owner named it. It is the difference between an app you can automate and an app you can only click, and it is what every other app in an estate integrates against: 49 apps publish a *.socket.yaml precisely because something else needs to call them.

SHC surface: the app publishes at least one *.socket.yaml, so SHC is itself calling the app’s API on another app’s behalf.

The owner named it, and it is the one capability that turns a catalogue of apps into a system: gitlab, glitchtip and wazuh ship a webhook.plug.yaml, hookshot publishes a webhook.socket.yaml, and nextcloud has a whole nextcloudWebhooks socket of its own. An app with no webhooks can only ever be a destination.

SHC surface: a webhook.plug.yaml, or a published socket whose name contains webhook.

THE self-hosting scale question, and it splits the catalogue right down the middle of comparable apps: nextcloud, immich-adjacent peertube, matrix, gitlab, plane and penpot can put their blobs in S3; paperless and immich cannot. It decides whether an app outgrows one disk.

SHC surface: an s3.plug.yaml in the app, or a published s3.socket.yaml. An app that SERVES object storage plainly supports it.

The owner named it, and for somebody choosing SELF-HOSTED software it is the anti-lock-in axis, the thing that makes leaving possible. 70 of 90 apps ship a backup.yaml, so SHC can already get the bytes out of most of them; whether the APP can hand you a portable document is a separate and less-answered question.

SHC surface: a backup.yaml in the app. When upstream has no export of its own, this is a SUBSTITUTE: SHC can still get your data out.

37 apps ship an *.exporter.yaml, but an exporter is frequently a sidecar translating an app that speaks no Prometheus at all, so the derived and declared answers genuinely disagree here. That disagreement is the reason the key is worth having: it tells an operator whether they are getting the app’s own instrumentation or our approximation of it.

SHC surface: an *.exporter.yaml in the app. When upstream exposes no metrics of its own, this is a SUBSTITUTE: the exporter is a sidecar SHC supplies.

The owner named it. It is the compliance gate: wazuh, midpoint and ciso-assistant exist to produce one, while kutt and shlink keep no record of who shortened what. SHC has no surface for it at all today, so it can only be declared, which makes it a second clean witness (with ldap) that the declared axis is not redundant with the derived one.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

The owner named it, and for the consumer-facing half of this catalogue it is the whole product: immich, nextcloud, jellyfin, vaultwarden and matrix are chosen BECAUSE the phone client exists, and superset, gitea and sonarqube will never have one. It is also the capability least visible from anything in this repository: no file here has ever mentioned an App Store listing.

SHC surface: none. This capability cannot be derived from any file in an app, because SHC has no integration for it. Declared or nothing.

Capabilitydeclared supportedpaidabsentunknownSHC wiresof which substitutes for an absent
sso4451247598
ldap100107n/an/a
mfa81099n/an/a
scim050103n/an/a
rbac013104120
multiuser40599n/an/a
multi_tenancy102105n/an/a
email_outbound430164470
email_inbound30010550
api90099650
webhooks80010050
object_storage140094220
import_export000108850
metrics000108360
audit_log012105n/an/a
mobile_app000108n/an/a

★ The gap list: what the app has and SHC does not wire

Section titled “★ The gap list: what the app has and SHC does not wire”

Every line is an app whose features: block says the application supports a capability, and whose files show SHC wiring nothing for it. This is SHC’s to-do list, in public. It is not a defect in the app and must never reduce the app’s score.

AppBucketCapabilityUpstream saysSHC hasNote
defectdojoenterprisessopaidnothingSAML/OIDC/OAuth SSO is DefectDojo Pro only as of 3.0. app.yaml: “the 3.x open-source build has no single si…
defectdojoenterpriserbacpaidnothingRBAC moved to Pro alongside SSO in 3.0.
defectdojoenterpriseapisupportednothingapp.yaml: “/api/v2/ stays reachable with a DefectDojo API token”.
docusealenterprisessopaidnothingapp.yaml: SSO/SAML is Pro at $240/year per user; “the free tier is uncapped on users but ships no SSO”.
fizzyintegratedwebhookssupportednothing
goalertflagshipwebhookssupportednothing
lookoutstandalonessosupportednothingKeycloak IS the app’s auth: better-auth fronts the tenant’s realm via the public lookout client. There is…
mailbox-reconcilerstandalonessosupportednothing
metabaseenterprisessopaidnothingapp.yaml: “Signing in with your company account (single sign-on over SAML, JWT or OIDC) is a paid Metabase …
planeenterpriseapisupportednothingCommunity ships the full v1 REST surface behind X-Api-Key personal tokens (apps/api/plane/api/ at v1.4.1): …
planeenterprisewebhookssupportednothingOUTBOUND workspace webhooks, on Community (app/urls/webhook.py at v1.4.1; WEBHOOK_ALLOWED_IPS/WEBHOOK_ALLOW…
shlinkstandaloneapisupportednothingThe REST API is the product; an API key is its only credential.
uptimeintegratedapisupportednothingAPI keys are issuable; app.yaml records that auth.disable “breaks API keys (issuing is disabled outright)”.
uptime-kumaflagshipapisupportednothingAPI keys are issuable; app.yaml records that auth.disable “breaks API keys (issuing is disabled outright)”.

14 gaps.

★ The reverse list: what SHC supplies and the app does not have

Section titled “★ The reverse list: what SHC supplies and the app does not have”

The other direction, which the owner asked for by name. An app here has an application that cannot do something, and an SHC integration that does it anyway: Keycloak in front of an app with no login of its own, a backup of an app that can export nothing, a sidecar exporter for an app that speaks no Prometheus. This is credit to the app, not to the application, and it is the reason a low upstream score does not have to mean a bad install.

AppBucketCapabilityUpstream saysSHC supplies
agent-canvasstandalonessoabsentaccess: gated (forward-auth)
onlyofficestandalonessoabsentaccess: gated (forward-auth)
openarchiverstandalonessoabsentaccess: gated (forward-auth)
openhandsstandalonessoabsentaccess: gated (forward-auth)
s3managerstandalonessoabsentaccess: gated (forward-auth)
stirling-pdfstandalonessoabsentaccess: gated (forward-auth)
uptimeintegratedssoabsentaccess: gated (forward-auth)
uptime-kumaflagshipssoabsentaccess: gated (forward-auth)

8 substitutions.

★ Inert integrations: a plug the application cannot consume

Section titled “★ Inert integrations: a plug the application cannot consume”

The failure mode neither axis finds alone, and the reason it is worth computing both. An app here ships a REAL in-app integration (an oidc plug, an smtp plug) against an application that has been MEASURED not to support the capability. Nothing is substituted and nothing works: the plug mints a Keycloak client nobody reads. These are defects in the app, and both of the ones below were already known and recorded in prose by the engineer who measured them; this table is the first time they are enumerated rather than buried.

AppBucketCapabilitySHC shipsWhy it cannot work
blockscoutstandalonessooidc.plug.yamlMeasured 2026-08-10 inside blockscout:9.0.2: ACCOUNT_KEYCLOAK_* appears nowhere and the only ueberauth stra…
monicastandalonessooidc.plug.yamlMeasured 2026-08-10 inside monica:4.1.2-apache and re-measured 2026-08-23: routes/ and config/ (auth.php in…

2 inert integrations.

Wired but undeclared: the cheapest work item on this page

Section titled “Wired but undeclared: the cheapest work item on this page”

SHC ships an integration and the app has not said whether the application supports the capability. These are not gaps and not substitutions; they are unknown values that somebody could close in seconds, because the integration working at all is strong evidence upstream supports it.

CapabilityAppsCount
ssoexternal-oidc, ffx, graphiti, multiplatform-keycloak, pulp5
rbacdependency-track, drover-relay, frappe, gitea, glitchtip, immich, nextcloud, paperless, peertube, pulp, superset, wazuh12
email_outboundexternal-smtp, ffx, fusionpbx, mail4
email_inboundexternal-imap, mailcow2
apiaffine-rs, clickhouse, crowdsec, dependency-track, docuseal, ethereum, external-gitlab, external-imap, external-llm, external-mailcow, external-mongodb, external-mysql, external-oidc, external-postgres, external-redis, external-s3, external-smtp, fleet, frappe, garage, gitlab, glitchtip, hockeypuck, hookshot, huly, hyperswitch, identity-broker, immich, ivatar, keygen, kubo, mail, mailcow, mariadb, matrix, mautic, metabase, midpoint, midpoint-classic, mongodb, monica, multiplatform-keycloak, mysql, netbird, nextcloud, nginx, ollama, onlyoffice, openreplay-ch, paperless, peertube, penpot, pmm, postiz, pulp, redis, s3, stirling-pdf, supabase, wazuh, windmill61
object_storageexternal-s3, fizzy, garage, glitchtip, openarchiver, s3, s3-migrate, supabase8
import_exportactivepieces, affine, agent-canvas, blockscout, ciso, ciso-assistant, clickhouse, crowdsec, defectdojo, defectdojo-classic, dependency-track, dify, docuseal, drover-relay, ethereum, external-s3, ffx, fizzy, fleet, frappe, fusionpbx, garage, gitea, gitlab, gitlab-runner, glitchtip, goalert, graphiti, greenbone, hockeypuck, hookshot, huly, huly-matrix-notifier, hyperswitch, immich, ivatar, jellyfin, journiv, keygen, kubo, kutt, librechat, mail, mailbox-reconciler, mailbox-reconciler-ffx, mailcow, mariadb, matomo, matrix, mautic, medusa, metabase, midpoint, midpoint-classic, mongodb, monica, mysql, netbird, nextcloud, nginx, ollama, onlyoffice, openarchiver, openhands, openreplay, openreplay-ch, paperless, peertube, penpot, plane, pmm, postiz, pulp, redis, shlink, sonarqube, stirling-pdf, supabase, superset, uptime, uptime-kuma, vaultwarden, wazuh, windmill, wordpress85
metricsblockscout, crowdsec, ethereum, ffx, fleet, frappe, gitea, graphiti, hockeypuck, hyperswitch, immich, keygen, kubo, mailcow, mariadb, matomo, matrix, mautic, metabase, mongodb, monica, mysql, netbird, nextcloud, nginx, paperless, peertube, penpot, postiz, pulp, redis, sonarqube, supabase, wazuh, windmill, wordpress36

213 wired-but-undeclared values.

license.id records what licence the code is under. It cannot say how OPEN the app is to somebody who wants to run it: an AGPL app can paywall SSO, and an SSPL app can be entirely free. license.openness is that second question, and every value in it is a human judgement: nothing in this repository can establish an upstream price list. The block therefore carries a reviewer and a date, and this table prints them.

ValueMeaning
fossEverything the operator runs is under an OSI-approved licence and there is no edition they cannot have.
open-coreThe code is open, but a paid edition holds capabilities the open one does not. NOT the same as requires_paid_license: fleet, sonarqube and ciso-assistant are open-core and the paid part (SCIM) is not something an estate must have.
proprietaryNo source, or source you may not run in production without paying.
source-availableSource is published; the licence restricts redistribution or resale but not self-hosting. SSPL, BUSL, Elastic-2.0, RSALv2, FCL. Costs a self-hoster nothing.
unknownNobody has established it. The default.
ValueMeaning
flatA paid tier priced per instance or per server. It does not get more expensive as the team grows.
freeNothing to pay for anything a self-hoster needs.
per-resourcePriced per host, agent, scan or document rather than per person: fleet per host, greenbone per instance-feed, docuseal $0.20/document via API.
per-user★ Priced per seat. docuseal is $240/yr/user on-prem; plane is $7/seat/mo; affine requires a subscription sized to the seat count.
unknownNobody has established it. The default.
AppSPDX idopenness.modelpricingSeat capPaid to runReviewed byOn
activepiecesMITopen-coreper-usernoneno[email protected]2026-08-24
affineunverifiedproprietaryper-user10yes[email protected]2026-08-18
affine-rsMITfossfreenoneno[email protected]2026-08-18
agent-canvasMITfossfreenoneno[email protected]2026-08-18
blockscoutunverifiedfossfreenoneno[email protected]2026-08-18
cisoAGPL-3.0-onlyopen-coreunknownnoneno[email protected]2026-08-18
ciso-assistantAGPL-3.0-onlyopen-coreunknownnoneno[email protected]2026-08-18
clickhouseApache-2.0fossfreenoneno[email protected]2026-08-18
crowdsecMITfossfreenoneno[email protected]2026-08-18
defectdojoBSD-3-Clauseopen-coreunknownnoneyes[email protected]2026-08-18
defectdojo-classicBSD-3-Clausefossfreenoneno[email protected]2026-08-18
dependency-trackApache-2.0fossfreenoneno[email protected]2026-08-18
difyLicenseRef-Dify-Modified-Apache-2.0source-availablefreenoneno[email protected]2026-08-18
docusealunverifiedopen-coreper-usernoneyes[email protected]2026-08-18
drover-relayMITfossfreenoneno[email protected]2026-09-02
ethereumGPL-3.0-or-later AND LGPL-3.0-or-laterfossfreenoneno[email protected]2026-08-18
fleetMIT AND LicenseRef-Fleet-Enterpriseopen-coreper-resourcenoneno[email protected]2026-08-18
frappeMIT AND GPL-3.0 AND AGPL-3.0 AND Apache-2.0fossfreenoneno[email protected]2026-08-18
fusionpbxMPL-1.1fossfreenoneno[email protected]2026-08-18
giteaMITfossfreenoneno[email protected]2026-08-18
gitlabMITopen-coreper-usernoneno[email protected]2026-08-18
gitlab-runnerMITfossfreenoneno[email protected]2026-08-18
glitchtipMITfossfreenoneno[email protected]2026-08-18
goalertApache-2.0fossfreenoneno[email protected]2026-08-21
graphitiApache-2.0fossfreenoneno[email protected]2026-08-18
greenboneunverifiedopen-coreper-resourcenoneno[email protected]2026-08-18
hockeypuckAGPL-3.0-onlyfossfreenoneno[email protected]2026-08-18
hookshotApache-2.0fossfreenoneno[email protected]2026-08-18
hulyEPL-2.0fossfreenoneno[email protected]2026-08-18
hyperswitchApache-2.0fossfreenoneno[email protected]2026-08-18
immichunverifiedfossfreenoneno[email protected]2026-08-18
ivatarunverifiedfossfreenoneno[email protected]2026-08-18
jellyfinunverifiedfossfreenoneno[email protected]2026-08-18
journivLicenseRef-Journiv-PolyForm-Noncommercial-with-Supplemental-Terms AND LicenseRef-Journiv-Web-Clientopen-coreunknownnoneno[email protected]2026-08-18
keygenLicenseRef-FCL-1.0-ALv2source-availablefreenoneno[email protected]2026-08-18
kuboMITfossfreenoneno[email protected]2026-08-18
kuttMITfossfreenoneno[email protected]2026-08-18
librechatMITfossfreenoneno[email protected]2026-08-18
mailMITfossfreenoneno[email protected]2026-08-18
mailcowunverifiedfossfreenoneno[email protected]2026-08-18
mariadbGPL-2.0-onlyfossfreenoneno[email protected]2026-08-18
matomoGPL-3.0-or-lateropen-coreflatnoneno[email protected]2026-08-18
matrixAGPL-3.0-or-laterfossfreenoneno[email protected]2026-08-18
mauticGPL-3.0-or-laterfossfreenoneno[email protected]2026-08-18
medusaMITfossfreenoneno[email protected]2026-08-18
metabaseAGPL-3.0open-coreper-usernoneyes[email protected]2026-08-18
midpointEUPL-1.2-or-laterfossfreenoneno[email protected]2026-08-18
midpoint-classicApache-2.0 OR EUPL-1.2fossfreenoneno[email protected]2026-08-18
mongodbSSPL-1.0source-availablefreenoneno[email protected]2026-08-18
monicaunverifiedfossfreenoneno[email protected]2026-08-18
mysqlGPL-2.0-onlyfossfreenoneno[email protected]2026-08-18
netbirdunverifiedfossfreenoneno[email protected]2026-08-18
nextcloudAGPL-3.0-onlyfossfreenoneno[email protected]2026-08-18
nginxBSD-2-Clausefossfreenoneno[email protected]2026-08-18
ollamaMITfossfreenoneno[email protected]2026-08-18
onlyofficeAGPL-3.0-onlyopen-sourcefreenoneno[email protected]2026-08-24
openarchiverAGPL-3.0-onlyopen-coreper-instancenoneno[email protected]2026-08-26
openhandsMITfossfreenoneno[email protected]2026-08-24
openreplayAGPL-3.0open-coreunknownnoneno[email protected]2026-08-18
paperlessunverifiedfossfreenoneno[email protected]2026-08-18
peertubeunverifiedfossfreenoneno[email protected]2026-08-18
penpotMPL-2.0fossfreenoneno[email protected]2026-08-18
planeAGPL-3.0-onlyopen-coreper-usernoneyes[email protected]2026-08-18
pmmAGPL-3.0-or-laterfossfreenoneno[email protected]2026-08-18
postizunverifiedfossfreenoneno[email protected]2026-08-18
pulpGPL-2.0-or-later AND Apache-2.0fossfreenoneno[email protected]2026-08-18
redisLicenseRef-RSAL-2.0 OR SSPL-1.0 OR AGPL-3.0source-availablefreenoneno[email protected]2026-08-18
s3managerApache-2.0fossfreenoneno[email protected]2026-08-18
shlinkMITfossfreenoneno[email protected]2026-08-18
sonarqubeLGPL-3.0-or-lateropen-coreflatnoneno[email protected]2026-08-18
stirling-pdfMITopen-coreper-user5no[email protected]2026-08-24
supabaseApache-2.0fossfreenoneno[email protected]2026-08-18
supersetApache-2.0fossfreenoneno[email protected]2026-08-18
uptimeMITfossfree1no[email protected]2026-08-18
uptime-kumaMITfossfree1no[email protected]2026-08-18
vaultwardenAGPL-3.0-onlyfossfreenoneno[email protected]2026-08-18
wazuhGPL-2.0-onlyfossfreenoneno[email protected]2026-08-18
windmillunverifiedopen-coreper-user10yes[email protected]2026-08-18
wordpressGPL-2.0-or-laterfossfreenoneno[email protected]2026-08-18

79 of 108 apps declare an openness block. The rest are unknown on both keys, which is the correct reading of silence and is not a mark against them.

✔ upstream supported · $ upstream paid-tier · ✘ upstream absent · ? unknown. A bold cell is one SHC wires; a ~ marks an SHC SUBSTITUTE for something upstream lacks.

Appssoldapmfascimrbacmultiusermulti_tenancyemail_outboundemail_inboundapiwebhooksobject_storageimport_exportmetricsaudit_logmobile_app
activepieces✔??????✔????~????
affine✔????✔?✔???✔~????
affine-rs✔??????????✔????
agent-canvas~✘???✘✘✘?????~????
blockscout✘??????✔?✔??~?~???
ciso✔??$????????~????
ciso-assistant✔??$????????~????
clickhouse????????????~????
crowdsec????????????~?~???
defectdojo$???$??✔?✔??~????
defectdojo-classic✔??????✔????~????
dependency-track✔???????????~????
dify??????✔?????~????
docuseal$????✔?✔????~????
drover-relay✔????✔✘?????~????
ethereum????????????~?~???
external-gitlab????????????????
external-imap????????????????
external-llm????????????????
external-mailcow????????????????
external-mongodb????????????????
external-mysql????????????????
external-oidc????????????????
external-postgres????????????????
external-redis????????????????
external-s3????????????~????
external-smtp????????????????
ffx????????????~?~???
fizzy✔??????✔?✔✔?~????
fleet✔?$$???✔????~?~???
frappe✔?✔????✔✔???~?~???
fusionpbx????????????~????
garage????????????~????
gitea✔??????✔????~?~???
gitlab✔??????✔??✔✔~????
gitlab-fizzy????????????????
gitlab-huly????????????????
gitlab-runner????????????~????
glitchtip✔??????✔??✔?~????
goalert✔??????✔?✔✔?~????
graphiti~????????????~?~???
greenbone???????✔????~????
hockeypuck???????✔????~?~???
hookshot??????????✔?~????
huly✔??????✔???✔~????
huly-matrix-notifier????????????~????
hyperswitch??✔?????????~?~???
identity-broker✔???????????????
immich✔??????✔????~?~???
ivatar✔??????✔????~????
jellyfin????????????~????
journiv✔???????????~????
keygen???????✔????~?~???
kubo????????????~?~???
kutt✔??????✔????~????
librechat✔?✔????✔???✔~????
lookout✔???????????????
mail????????????~????
mailbox-reconciler✔???????????~????
mailbox-reconciler-ffx????????????~????
mailcow✔??????✔????~?~???
mariadb????????????~?~???
matomo✔ᵖ??????✔????~?~???
matrix✔??????✔???✔~?~???
mautic✔??????✔✔???~?~???
medusa???????????✔~????
metabase$✔?????✔????~?~???
midpoint✔???????????~????
midpoint-classic✔???????????~????
mongodb????????????~?~???
monica✘??????✔????~?~???
monica-nextcloud✘???????????????
multiplatform✔???????????????
multiplatform-keycloak????????????????
mysql????????????~?~???
netbird✔???????????~?~???
nextcloud✔?✔ᵖ????✔✔?✔✔~?~???
nginx????????????~?~???
ollama????????????~????
onlyoffice~✘???????????~????
openarchiver~✘?✔??✔?✘????~????
openhands~✘???????????~????
openreplay$ᶠ??????✔???✔~????
openreplay-ch????????????~????
paperless✔??????✔????~?~???
peertube✔??????✔???✔~?~???
penpot✔??????✔???✔~?~???
plane$??????✔?✔✔✔~????
pmm????????????~????
postiz✔??????✔????~?~???
pulp???????????✔~?~???
redis????????????~?~???
s3????????????????
s3-migrate????????????????
s3manager~✘????✘?????✔????
shlink✘????✘?✔?✔??~????
sonarqube✔??$???✔?✔??~?~?$?
static-site????????????????
stirling-pdf~✘???????????~????
supabase✔??????✔????~?~???
superset✔???????????~????
uptime~✘?✔?✘✘?✔?✔??~??✘?
uptime-kuma~✘?✔?✘✘?✔?✔??~??✘?
vaultwarden✔?✔????✔????~????
wazuh✔??????✔??✔?~?~???
whoami????????????????
windmill✔??$???✔????~?~???
wordpress✔??????✔????~?~???

A features: entry off the vocabulary is dropped, silently at the render surface, loudly here and on this generator’s stderr. A bad STATE token degrades to unknown, never to absent: a typo must not be able to assert that a capability is missing.

None. Every declaration in the tree is in the vocabulary.