Linear provider

The Linear provider exposes teams and issues as paths. Issue listings can preload field files, so reading a listed issue’s title, state, priority, or assignee can avoid another upstream call when the cached object is warm. description.md is loaded from the single issue object, not from the list query.

At a glance

Field	Value
Provider id	linear
Default mount	/linear
WASM	omnifs_provider_linear.wasm
Manifest	providers/linear/omnifs.provider.json
Routes	providers/linear/src/lib.rs
Auth	Linear OAuth by default; personal access token also supported
Capabilities	api.linear.app for GraphQL reads; declared memory need 128 MB

Setup

omnifs init linear
omnifs up
omnifs shell

Linear defaults to OAuth and also supports a personal access token scheme. The host owns the credential and injects the approved Authorization header for api.linear.app.

Main paths

/linear/teams
/linear/teams/{team}/issues
/linear/teams/{team}/issues/{filter}
/linear/teams/{team}/issues/{filter}/{ident}
/linear/teams/{team}/issues/{filter}/{ident}/item.json
/linear/teams/{team}/issues/{filter}/{ident}/item.md
/linear/teams/{team}/issues/{filter}/{ident}/title
/linear/teams/{team}/issues/{filter}/{ident}/state
/linear/teams/{team}/issues/{filter}/{ident}/priority
/linear/teams/{team}/issues/{filter}/{ident}/assignee
/linear/teams/{team}/issues/{filter}/{ident}/description.md

{team} is the Linear team key, such as ENG. {filter} is a provider-supported issue filter such as open or all. {ident} is the full issue identifier, such as ENG-1421.

Examples

ls /omnifs/linear/teams
ls /omnifs/linear/teams/ENG/issues/open
cat /omnifs/linear/teams/ENG/issues/open/ENG-1421/title
cat /omnifs/linear/teams/ENG/issues/open/ENG-1421/item.md
cat /omnifs/linear/teams/ENG/issues/open/ENG-1421/state
cat /omnifs/linear/teams/ENG/issues/open/ENG-1421/description.md

Capability notes

The provider declares api.linear.app, expected memory needs, and auth injection. GraphQL requests are still host-mediated callouts.

Limitations

Issue lists can truncate after a large result set and remain open to direct lookup. Linear GraphQL single-issue reads do not use HTTP ETag revalidation, so expect upstream fetches when the provider needs a fresh full issue.