doctor

Run the environment-diagnostic checklist from the rn-setup skill. Walk all 17 prerequisite checks (Node.js version, CDP bridge dependencies, rn-fast-runner build (iOS), rn-android-runner build/install (Android), maestro-runner, iOS simulator, Android emulator, Metro dev server, CDP connection, injected __RN_AGENT helpers, ffmpeg, idb (screen-mirror fast path), physical-device prerequisites, plugin version freshness, Vercel rules sync freshness, CDP auto-reconnect mode, linked-worktree action inheritance) and surface install commands for any missing dependencies.

iOS device automation is owned by the in-tree rn-fast-runner XCTest project (D1219, PR #164); Android device automation is owned by the in-tree rn-android-runner (UiAutomator instrumentation). These in-tree runners are the sole device backend — there is no external CLI to install. Mark rn-android-runner as N/A on iOS-only setups and rn-fast-runner as N/A on Android-only / non-macOS setups. If a device tool fails with RUNNER_PROTOCOL_MISMATCH, the installed/prebuilt runner artifact predates the plugin's wire protocol: on iOS delete packages/rn-fast-runner/build/DerivedData and re-open the device session (or re-run xcodebuild build-for-testing); on Android re-run ./gradlew :app:assembleDebug :app:assembleDebugAndroidTest and adb install -r both APKs.

This command is read-only. It diagnoses the current environment and recommends fixes. It does NOT modify any files in the user's project, inject documentation, or instrument source code.

Present results as a 17-row table. Follow the detailed probes and remediation rules in rn-setup; do not infer runner provenance, helper availability, or auto-reconnect configuration from passive cdp_status.

For iOS replay compatibility, also run the packaged maestro-runner-pin.js diagnose --json report. State the proof-domain split: exact React testIDs use react-tree; native/system selectors use xctest-native. A runtime-version heuristic is diagnostic context only and never proof of blindness. Report NATIVE_SURFACE_BLIND only when the replay includes the bounded same-screen native-selector evidence; otherwise preserve the ordinary selector miss. The supported next action is the central native WDA smoke on a WDA-healthy runtime.

For linked-worktree action inheritance, resolve the RN app root and run the packaged worktree-inheritance.js plan --host claude --app-root <app> --json. Report .rn-agent/actions as tracked, inherited, missing, explicit-legacy-repair-needed, unsafe, or refused. Also report hook status. Never apply, repair, install a hook, print a private source path, or inspect an action body from doctor.

For idb, require both a healthy idb --help and idb_companion; an active install PID is INSTALLING. A client on PATH whose idb --help fails is BROKEN, not MISSING — report it as installed but not responding, naming the Python 3.14 interpreter incompatibility as the probable cause rather than a certainty (the probe cannot separate a crash from a timeout or EACCES), and give the pinned repair command, never a bare pipx install fb-idb. On a machine with no supported interpreter, prefix that command with brew install [email protected] &&. Otherwise report MISSING and the documented install command. This row is YELLOW because mirroring has a screenshot fallback.

For runner provenance, inspect the runner artifact/state metadata documented by rn-setup. For helpers, use a narrow gated CDP read. For CDP auto-reconnect, resolve RN_CDP_AUTOCONNECT over .rn-agent/config.json, then the default.

For plugin-version or Vercel-rule drift, report the documented command but do not execute it. Offline version checks do not fail the plugin.

For session-authority health, run the packaged read-only probe from the RN app root: node "${CLAUDE_PLUGIN_ROOT:-${RN_DEV_AGENT_CODEX_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:?set it to the installed rn-dev-agent plugin root, then re-run}}}/rn-dev-agent-core/dist/session-doctor.js" report --json. The report is three-state: GREEN when sameRootOwner is absent and wedged and repairable are both false, YELLOW when repairable is true (a proven-dead owner of this exact root, which the next transport start or the repair command releases on its own) or sameRootOwner is live (another session holds this worktree right now — report its ownerSession and ownerAppRoot instead of GREEN; it is never released and clears when that session closes), and RED when wedged is true. On RED, name the exact cause it returns rather than a generic stale-lock story: sameRootOwner: unprovable (the recorded owner's process identity cannot be read, so it is conservatively treated as live), startupCleanupBlocked (the owner is proven dead but an obligation such as RUNNER_ADOPTION_REQUIRED could not be discharged), or ownerIsThisRoot: false, whose ownerMismatch says which: app-root (a proven-dead owner of a different app root in this worktree) or source-identity (the same app root under different declared manifests). Also report abandonedContenders when it is non-zero. The supported repair is node "${CLAUDE_PLUGIN_ROOT:-${RN_DEV_AGENT_CODEX_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:?set it to the installed rn-dev-agent plugin root, then re-run}}}/rn-dev-agent-core/dist/session-doctor.js" repair, which runs the same proven-dead startup cleanup a fresh transport runs; print it for the user to run, but do not execute it from doctor. Print it rooted where it can succeed: for ownerMismatch: app-root that is the reported ownerAppRoot, because repair from the current root can never release another root's owner. For ownerMismatch: source-identity do not re-root it — the payload's own remedy names the declared-manifest restore that makes this root's repair work. It never releases a live or unprovable owner and there is no force-steal — do not suggest deleting or moving files in the authority store. Not every stale lock self-heals: only a proven-dead owner does.