omnifs v0.2.1

Fixed

• omnifs up now passes the runtime container's config, cache, mount-config, and provider directories through OMNIFS_* environment variables. This keeps the startup readiness check aligned with the materialized session mounts, so release containers no longer report zero providers after the FUSE mount is already live.

omnifs v0.2.0

Added

• Object-shaped provider routing and host-owned object/view caching now back the current provider surfaces. Providers can store canonical upstream bytes and render multiple filesystem leaves from the same object, so repeated reads of related fields can be served from host cache instead of refetching the upstream service.
• ~/.omnifs/config.toml is now the normal user-authored configuration surface for mounts. omnifs init writes inline [[mounts]] entries, and mount discovery still reads legacy per-mount JSON files so existing setups continue to load.
• The public README is rewritten for launch around the shipped alpha surface: six live providers, the npm + runtime-container install path, Linux FUSE caveat, host-mediated provider authority, and the read-only roadmap boundary.

Changed

• Provider paths now use plain resource names instead of underscore-prefixed control names. GitHub exposes repo, issues, pulls, and actions; filters are open and all; DNS exposes resolvers, reverse, all, and raw; Docker exposes /containers.json, /compose.json, and {by-name,by-id,running,stopped} under /containers.
• Built-in providers have been rewritten against the current SDK registration model, removing migration-era provider boilerplate while preserving the shipped GitHub, DNS, arXiv, Docker, Linear, and SQLite path surfaces.
• The default credential backend is now the private credentials.json file. The keychain backend remains available through OMNIFS_CREDS_BACKEND=keychain, but normal startup, auth, and runtime refresh paths no longer trigger platform keychain prompts by default.

Fixed

• CI host-test Wasmtime cache restore now uses stable OS/architecture restore keys with per-run write keys, improving cache reuse between push and pull-request runs without depending on synthetic merge SHAs.

omnifs v0.2.0-dev.2

Added

• Container shell now greets interactive users with a welcome banner: an "OMNIFS" wordmark, the tagline open a path, read the world., and indented blocks of example paths (ls /github/<owner>/<repo>/_repo, cat /dns/<domain>/TXT, an arXiv find pipeline) and useful commands (omnifs status, omnifs logs -f, omnifs auth list). Gated on [[ -o interactive ]] so zsh -c '...' invocations stay silent. Lives in scripts/container-zshrc.zsh, copied into both the dev Dockerfile and the release scripts/ci/Dockerfile.runtime.
• omnifs up now prints a hint pointing at omnifs shell after the FUSE mount comes online, so new users immediately know how to enter the projected filesystem.
• omnifs inspect shows a live JSONL observability stream from the host daemon as a ratatui TUI: a path-tree of mount activity, a per-mount sparkline strip, an operations log (retention 4096), and an inline waterfall for the selected trace. Use --replay <file> for a captured JSONL trace, --record <file> to tee the live stream to a host path while attached, and --plain for line-oriented output. The host daemon emits typed InspectorEvent records (FUSE, provider, callout, subtree, clone, cache) through a non-blocking sink (lock-free crossbeam_queue::ArrayQueue history ring plus tokio::sync::broadcast for live subscribers) over TCP loopback 127.0.0.1:7878. Schema lives in the new omnifs-inspector crate with redaction at the wire boundary. File tee is opt-in via OMNIFS_INSPECTOR_PATH. omnifs up exposes the inspector port to host loopback so omnifs inspect works against the standard runtime container, not just omnifs dev.
• Runtime images now carry an ai.0xff.omnifs.min-launcher-version label, and omnifs dev refuses to start an image that requires a newer launcher. This catches source-image versus installed-CLI skew before Docker starts a container with missing port mappings or environment wiring.
• OMNIFS_CREDS_BACKEND=file|keychain overrides the automatic credential backend selection, and omnifs dev uses the file backend directly so contributor builds do not block on macOS Keychain prompts from differently signed binaries.

Changed

• Repository moved from raulk/omnifs to 0xff-ai/omnifs. All repository URLs, GHCR image references (ghcr.io/0xff-ai/omnifs), npm package homepage/repository metadata, README and documentation links, and the omnifs dev clone hint are updated. The OMNIFS_DEMO_OWNER default in scripts/demo.sh now points to 0xff-ai. npm package names are unchanged (@0xff-ai/omnifs and @0xff-ai/omnifs-cli-*).
• omnifs setup provider multiselect now sorts db and linear to the bottom and starts them unchecked. Both require user-supplied state (a SQLite fixture path, a Linear API key) that the smoke onboarding flow can't satisfy from ambient context.

Fixed

• omnifs inspect now reports honest socket state. The TUI starts in a waiting state until the daemon actually accepts a TCP connection, and --plain mode prints connect, disconnect, and delayed "no inspector listening" diagnostics instead of silently reconnecting forever.
• GitHub repository names such as .github and .gitignore are accepted as safe path segments. Only bare . and .. remain rejected for traversal safety.
• GitHub issue listings under a nonexistent repository now surface NotFound instead of InvalidInput, so FUSE renders ENOENT rather than EINVAL for structurally valid paths that point at a missing repo.
• GitHub OAuth device flow no longer fails on the first poll with request_failed: Failed to parse server response. GitHub's /login/oauth/access_token returns 200 OK with {"error":"authorization_pending",...} while the user is approving, violating RFC 8628; oauth2 5.x interpreted the 200 as success and tripped on the body schema. omnifs-auth now wraps reqwest::Client in a DevicePollingHttp impl of oauth2::AsyncHttpClient that re-stamps any 200 + JSON-with-error-field response as 400, routing GitHub's responses through the standard error-handling path. Compliant providers are unaffected.
• omnifs dev builds again. The Dockerfile still referenced .cargo/, which was removed in #78; the stray COPY .cargo .cargo and matching .dockerignore allowlist entries are gone.

omnifs v0.2.0-dev.1

Fixed

• omnifs CLI shim no longer fails at startup with Cannot find module '../../platforms.json'. The platform-to-package map was previously loaded at runtime from npm/platforms.json, which lives outside the published @0xff-ai/omnifs tarball and is not included on install. The map is now inlined in scripts/resolve-binary.js, with just npm-validate cross-checking it against npm/platforms.json so the two cannot drift.

omnifs v0.2.0-dev.0

Added

• omnifs is now distributed on npm as @0xff-ai/omnifs. Install with npm install -g @0xff-ai/omnifs; the CLI binary ships in one of four platform-specific optional-dependency packages (darwin-arm64, darwin-x64, linux-arm64, linux-x64). The Docker image is pulled on omnifs up, not at install time.
• Full omnifs init <provider> / omnifs up / omnifs down mount lifecycle. omnifs init walks through provider auth (device-code or PKCE loopback OAuth), writes a per-mount config under ~/.omnifs/config/mounts/, and stores credentials in the OS keychain (macOS Keychain, Linux libsecret, Windows DPAPI) with a mode-600 file fallback at ~/.omnifs/data/credentials.json. omnifs up pulls the matching runtime image, materialises credentials into a private session directory bind-mounted read-only into the container, then removes the session on stop or failure.
• omnifs auth subcommands: login, logout, status, refresh, scopes, import. OAuth refresh happens automatically with a one-shot 401 retry coordinated by a singleflight plus a cross-process file lock so concurrent CLI invocations do not race on the same refresh token.
• omnifs setup guided first-run walkthrough: detects OS and Docker, helps pick providers, runs init for each, and brings the container up.
• omnifs doctor runs ten ordered probes to diagnose why a mount is not working (Docker availability, FUSE timeout, missing credentials, etc.).
• omnifs reset clears configs and credentials after an explicit confirmation prompt.
• omnifs mounts ls / omnifs mounts rm for listing and removing configured mounts.
• omnifs status readiness card showing runtime state, configured mounts, and auth state.
• omnifs version and omnifs completions commands.
• omnifs dev contributor sandbox: walks up from cwd to find the workspace Cargo.toml, captures gh auth token, downloads the Chinook SQLite fixture into .secrets/db/test.db, builds an omnifs:<short-sha>-dev image, and starts a container with all built-in providers mounted. Replaces the old just dev / docker compose up workflow.
• Three supported OAuth flows in the new omnifs-auth crate: PKCE loopback (browser-redirect), PKCE manual code (paste-back), and device code (visit URL, type short code). GitHub uses device code with no default write scopes; Linear uses PKCE loopback with the read scope.
• Host-managed provider credentials: providers never see tokens; the host attaches them to outgoing HTTP requests after callouts cross the WASM boundary. Provider auth needs are declared in omnifs.provider.json (OAuth endpoints, scopes, injection header, allowed domains); adding a new service does not require patching the host.
• omnifs-creds crate for the keychain + file + in-memory credential store, with CredentialKey::storage_key() (provider:scheme:account) as the public wire form. Stale file-fallback entries are cleaned up after successful keyring writes; durability and permissions are hardened.
• Database provider (omnifs-provider-db) mounted at /db, projecting a read-only SQLite database as a filesystem. Exposes meta/{version.txt,path.txt,info.json} and tables/{name}/{schema.sql,schema.json,indexes.json,count.txt,sample.json}. SQLite runs inside the WASM sandbox via rusqlite with the bundled feature; the host preopens the database file's parent directory through Wasmtime's WASI context with read-only permissions so no bytes cross the WIT boundary. v1 is SQLite-only and read-only.
• Docker provider (omnifs-provider-docker) mounted at /docker, projecting the local Docker daemon over the Unix socket. Exposes /system/{info,version,df}.json, /system/ping, /containers/_listing.json, facets {by-name,by-id,_running,_stopped} each binding to a per-container subtree (inspect.json, summary.json, summary.txt, state), and /compose/{project}/services/{service}/containers/{name} grouping by Compose labels. Container state files are marked volatile so reads bypass the kernel page cache and always reach the provider. Bounded-window /events polling translates container actions into cache-invalidation prefixes.
• Unix-socket HTTP transport in the host. Providers use HttpEndpoint::Unix and build_url(path, query); the host detects unix: URLs, decodes the hex-encoded socket path, builds a per-socket reqwest::Client via ClientBuilder::unix_socket, and rewrites the URL for that client. Unix sockets are gated by a new unix-sockets allowlist on CapabilityGrants.
• Linear provider (omnifs-provider-linear) mounted at /linear, projecting a Linear workspace. Teams appear at /linear/teams/{KEY}, issues at /linear/teams/{KEY}/issues/{_open,_all}/{IDENT}/. Each issue surface has title, state, priority, assignee, and description.md as files. Issue listings preload child files so a cat after ls skips a follow-up round trip. Uses Linear's GraphQL API with hand-written query strings and serde response structs (Linear's endpoint rejects full introspection queries as too complex, ruling out code generation).
• arXiv provider now also exposes a recent-submissions surface per category: /categories/{cat}/recent, /categories/{cat}/recent/_fetched, /categories/{cat}/recent/pages/{n}, and /categories/{cat}/submissions/{YYYYMMDD} directories discovered from fetched pages. Direct paper lookup at /papers/{id} is unchanged.
• Projected file attributes: providers now declare Size (Exact, NonZero, or Unknown), Bytes (inline or deferred), ReadMode, and Stability (Immutable, Mutable, or Volatile) through the Projection API. The host uses these facts to set st_size, FUSE direct-I/O flags, cache behavior, ranged-read handling, and post-read size promotion. The old 256 MiB placeholder is removed. Volatile files return entry_timeout = 0, attr_timeout = 0, and FOPEN_DIRECT_IO; ranged files open a provider handle for snapshot-consistent reads.
• Sandboxed archive extraction via a host-owned omnifs-tool-archive Wasm component. BlobExecutor streams fetch-blob responses into a staged temp file and commits metadata before the body is visible; ArchiveExecutor keys extracted trees by (cache-key, format, strip-prefix) and coalesces concurrent extractions through TreeMaterializer. The extractor runs path sanitization, depth/length/entry-count/per-file/total-byte limits inside the sandbox and publishes completed trees via atomic directory rename so tree refs never observe partial output.
• EffectiveConfig type representing a mount after provider metadata has been merged in. ProviderCatalog::load_mount() returns one; credential targeting, session materialisation, and runtime construction all consume it. The previous InstanceConfig-plus-late-apply_metadata pattern is removed.
• Provider runtime capabilities now come back from init as (State, ProviderInfo, RequestedCapabilities) instead of a separate WIT export. Initialisation runs exactly once in ProviderRuntime::new. Capability entries can be marked dynamic: true when the concrete grant depends on mount config (Docker's socket path is the motivating case).
• omnifs-mount-schema crate is split into typed modules with a checked-in JSON schema at crates/omnifs-mount-schema/schema/omnifs.provider.schema.json (regenerate with just regen-schema).
• Per-crate README files for all published crates.

Changed

• arXiv provider route model is restructured around recent submissions. The calendar/date-query, new, updated, by-author, /authors, and /search surfaces are removed. Category traversal now goes through /categories/{category}/recent and /categories/{category}/submissions/{YYYYMMDD}; direct paper lookup at /papers/{paper} is unchanged. The only live category listing query shape is search_query=cat:{category} sorted by submittedDate descending.
• Provider protocol vocabulary is reorganized into three orthogonal channels: callout (intermediate host work the provider suspends on), return (the completed operation answer), and effect (host-side mutation committed at the return boundary). provider-response is renamed to provider-step with arms suspended(callouts) and returned(provider-return). A return cannot carry callouts; an error return cannot carry effects.
• Dead git callouts (git-list-tree, git-read-blob, git-head-ref, git-list-cached-repos) and the unused reconcile interface are removed from the WIT.
• The sidecar::materialize method is folded into lookup-child and list-children via #[subtree] dispatch from the SDK.
• Host runtime module crates/host/src/runtime/mod.rs is split into focused modules: instance.rs (Wasmtime mechanics), callouts.rs (dispatch and tracing), effects.rs (terminal mutations and ProjectionAccumulator), log_redaction.rs, wit_conversions.rs, op.rs (the Op enum and Validator), and http_stack.rs (shared HTTP transport). RuntimeError construction errors split into RuntimeBuildError.
• Browse cache re-skin enums collapse into their WIT counterparts; cache::SCHEMA_VERSION bumps to 5, invalidating existing L2 records.
• CLI flows are redesigned around mount configs, provider metadata, credential materialisation, and container lifecycle commands. Verbose output is off by default; -v enables INFO logs and -vv adds DEBUG. Common errors surface a Try: block with a concrete next step.
• Provider manifests (omnifs.provider.json) describe auth schemes, token injection policy, capability grants, and config schema. All built-in providers (arxiv, db, dns, docker, github, linear, test) carry an omnifs.provider.json and drop vestigial [package.metadata.component] Cargo sections.
• Docker Compose development entrypoints (compose.yaml, just dev) are replaced by the supported omnifs dev, omnifs shell, omnifs logs, and omnifs down workflow.
• Wasm providers are baked into the runtime image at `/root/.om...

0.1.0 - 2026-05-07

Release Notes

Added

• FUSE filesystem on Linux that projects external services into local paths, with macOS and Windows planned.
• WASM-based provider architecture using the WIT Component Model (wasmtime); each provider is a wasm32-wasip2 component implementing the omnifs:provider WIT interface.
• Per-provider capability declarations (HTTP domains, auth types, memory limits, git/websocket/streaming flags) enforced by the host runtime.
• GitHub provider mounted at /github, projecting repos, issues, PRs, CI runs, diffs, and source trees as files. Source trees are bind-mounted clones (cloned on demand via SSH); issues and PRs project per-item directories with title, body, state, and comments as separate files.
• Git-backed reconciliation scaffolding via a custom remote helper (read path live; mutation path WIP per the design docs).
• arXiv provider mounted at /arxiv, projecting per-paper subtrees (paper.pdf, source.tar.gz, metadata.json, links.json, versions/v{n}/) under /arxiv/papers/{id} and the /categories/{cat}/{ym|new|updated|by-author}, /authors/{author}/{...|by-category}, and /search/{query} scopes.
• DNS provider mounted at /dns, projecting record types (A/AAAA/MX/NS/TXT/CNAME/SOA/SRV plus _all and _raw) over DNS-over-HTTPS, with resolver-scoped queries via /dns/@{resolver}/... and reverse lookups via /dns/_reverse/{ip}.
• Path-first SDK with attribute macros: #[dir], #[file], #[treeref], #[bind], #[mutate] inside #[handlers] impl ... blocks; #[subtree] impl B { ... } for typed subtree dispatch (mounted at any number of #[bind("...")] sites); #[config] and #[provider(mounts(...))] at the top level. Path patterns support literal segments, bare captures ({name}), prefix captures (v{ver}), and rest captures ({*tail}).
• Auto-navigable intermediate directories: any literal-segment prefix of a registered route is a directory without an explicit handler, so providers don't write no-op stubs for navigation nodes. Listings carry exhaustive=false whenever a sibling capture or rest segment lives at depth+1, preventing the host's negative cache from short-circuiting valid dynamic-capture lookups.
• Parse-function fallthrough in route matching: when the highest-precedence pattern's parse function rejects a candidate path, the dispatcher falls through to the next-most-specific candidate. Per-segment validators participate in match candidacy rather than acting as a post-match check.
• Two-tier (L0 in-memory, L2 redb-backed) browse cache with negative caching, projected-file extraction from listings, and event-outcome driven invalidation; sibling preload on lookup terminals so the host caches projected siblings alongside the primary entry.
• Cross-listing PR preload and hybrid issue/PR pagination in the GitHub provider.
• HTTP SDK gains POST + raw / JSON bodies and adopts http crate request/response types.
• Docker Compose dev workflow (compose.yaml, just dev, just shell) with self-starting published image.
• docs/design/path-dispatch-and-listing.md as the perennial reference for routing precedence, listing semantics, and the lookup/readdir authority split; docs/design/projected-file-sizes.md documenting the direct_io redesign.

Changed

• SDK and host runtime are redesigned around path-first handlers and callouts. Effect-based handler signatures are replaced by free-function path handlers that return either a terminal op-result or a list of callouts for the host to execute and resume. The previous effect-style API is removed.
• Provider configuration moved from TOML to JSON. The host parses each mount's JSON config into InstanceConfig and re-serializes the provider-specific "config" object as JSON bytes for initialize(); providers deserialize via serde_json::from_slice (the SDK's #[config] macro wires this up automatically).
• Subtree handoff is now declared with #[treeref("...")] (renamed from the previous #[subtree]) so #[subtree] is free for typed-subtree-dispatch impl blocks. The WIT result variant is still spelled subtree.
• Dir and file handlers can co-exist on identical rest-captured templates; the parent dir handler's projection authoritatively decides the child's kind at lookup.
• Handler cx parameter is now optional, and the DirCx lifetime is dropped.

Fixed

• Bind exact-match lookups no longer return a bare Lookup::entry with an implied exhaustive: true over an empty sibling set. The host's lookup-side cache treated the bare entry as "the bind has no children" and wrote an exhaustive empty Dirents at the bind site, causing subsequent readdir to short-circuit before the typed subtree's list_children ran.
• Parent listings that returned exhaustive empty no longer poison dynamic-capture child lookups. With the auto-navigable rule and the SDK's exhaustive-flag computation, listings under parents with capture children at depth+1 are correctly non-exhaustive, so the host's negative cache leaves room for the capture handler to dispatch.
• GitHub route projections normalized to honor sibling preload, projected-files, and cross-listing PR preload paths.

Download omnifs-cli 0.1.0

File	Platform	Checksum
omnifs-cli-aarch64-unknown-linux-gnu.tar.xz	ARM64 Linux	checksum
omnifs-cli-x86_64-unknown-linux-gnu.tar.xz	x64 Linux	checksum

Verifying GitHub Artifact Attestations

The artifacts in this release have attestations generated with GitHub Artifact Attestations. These can be verified by using the GitHub CLI:

gh attestation verify <file-path of downloaded artifact> --repo raulk/omnifs

You can also download the attestation from GitHub and verify against that directly:

gh attestation verify <file-path of downloaded artifact> --bundle <file-path of downloaded attestation>

Demo assets

GIF and MP4 for README. Not a software release.