Add structured automation outputs and bump to 0.1.35-rc.2

README.md

@@ -109,13 +109,15 @@ Published package notes:
 ## Commands
 
 ```bash
+rdt doctor --session demo
 rdt session open --url http://localhost:3000 --browser chromium --engine auto --session demo
 rdt session connect --ws-endpoint ws://127.0.0.1:3000/ --target-url localhost:3000 --session remote
 rdt session attach --cdp-url http://127.0.0.1:9222 --target-url localhost:3000 --session cdp
 rdt session doctor --session demo
 rdt tree get --session demo
+rdt tree stats --session demo --top 5
 rdt interact type --session demo --selector 'input[name="query"]' --text hello
-rdt node search App --session demo --snapshot <snapshotId>
+rdt node search App --session demo --snapshot <snapshotId> --structured
 rdt node inspect <nodeId> --session demo --snapshot <snapshotId>
 rdt profiler start --session demo
 rdt profiler stop --session demo
@@ -126,6 +128,7 @@ rdt profiler export --session demo --compress
 ## Snapshot Semantics
 
 - `tree get` returns a `snapshotId`.
+- `tree stats` returns the same `snapshotId` plus summary fields such as `rootCount`, `nodeCount`, `rootSummaries`, and `topLevelComponents`.
 - Node IDs are only meaningful within that snapshot.
 - The runtime currently keeps up to `5` snapshots in memory per session.
 - Agent-friendly recommended flow:
@@ -139,6 +142,7 @@ rdt profiler export --session demo --compress
 ## Doctor
 
 - `rdt session doctor --session <name>` reports runtime readiness and trust boundaries before a deeper investigation.
+- `rdt doctor --session <name>` is a first-class alias for the same command.
 - It also reports `enginePreference`, `selectedEngine`, `recommendedEngine`, `availableEngines`, and DevTools capability hints so agents know whether they are on a custom fallback or a DevTools-aligned path.
 - `sourceCapability` is reported separately from engine selection.
 - `_debugSource` is treated as an optional legacy source-mapping capability, not an engine-selection gate.
@@ -154,12 +158,15 @@ rdt profiler export --session demo --compress
 
 - Use built-in `interact` commands before reaching for external Playwright helper scripts.
 - Current supported actions:
-  - `rdt interact click --session <name> --selector <css>`
+  - `rdt interact click --session <name> --selector <css> [--delivery auto|playwright|dom]`
   - `rdt interact type --session <name> --selector <css> --text <value>`
   - `rdt interact press --session <name> --key <name> [--selector <css>]`
   - `rdt interact wait --session <name> --ms <n>`
 - These commands execute through the same Playwright session that owns the current `rdt` browser page.
 - They target the first matching selector only and return structured action metadata plus trust-boundary fields.
+- `interact click` defaults to `--delivery auto`.
+- In `auto`, profiler-active clicks fall back to DOM dispatch and report `requestedDelivery`, `effectiveDelivery`, `profilerActive`, and `fallbackApplied`.
+- Use `--delivery playwright` to force Playwright pointer input, or `--delivery dom` to force DOM dispatch.
 - `click`, `type`, and `press` confirm that the action was dispatched. They do not guarantee that the page or React tree has fully settled afterward.
 - When profiling or triggering large rerenders, follow `interact` with an explicit verification step such as `interact wait`, `tree get`, `node inspect`, or a profiler read command.
 
@@ -168,10 +175,10 @@ Example deterministic flow:
 ```bash
 rdt tree get --session demo
 # => save snapshotId from output
-rdt node search App --session demo --snapshot <snapshotId>
+rdt node search App --session demo --snapshot <snapshotId> --structured
 rdt node inspect <nodeId> --session demo --snapshot <snapshotId>
 rdt node highlight <nodeId> --session demo --snapshot <snapshotId>
-rdt source reveal <nodeId> --session demo --snapshot <snapshotId>
+rdt source reveal <nodeId> --session demo --snapshot <snapshotId> --structured
 ```
 
 Snapshot recovery:
@@ -269,7 +276,11 @@ Use `node pick` when the agent knows the visible element but not the component n
 - `hooks` is a simplified serialized view of hook state from the inspected fiber.
 - `context` is a serialized view of current context dependencies for the inspected node.
 - `source` projects `_debugSource` when available; `null` is expected in many dev builds.
+- `source reveal --structured` returns `status`, `available`, `mode`, `reason`, and `source` so automation can distinguish unavailable source data from a successful source payload.
+- `source reveal` without `--structured` preserves the raw legacy behavior and may return literal `null`.
 - `dom` is the first host element summary used for CLI highlight and DOM-oriented inspection.
+- `node search --structured` wraps search results in `{ items, query, snapshotId, matchCount, runtimeWarnings }`.
+- When `node search --structured` returns `matchCount: 0`, `runtimeWarnings` explains that the component may be absent from the current snapshot rather than absent from the codebase.
 - Profiler summary fields are commit-oriented CLI metrics, not the full DevTools profiler session schema.
 - `profiler summary` and exported summaries explicitly report:
   - `measurementMode: "actual-duration" | "structural-only" | "mixed"`

docs/devtools-concept-mapping.md

@@ -34,16 +34,18 @@ It is a maintenance aid for agents and contributors. It is not a commitment to r
 | `__REACT_DEVTOOLS_GLOBAL_HOOK__` shim | `src/runtime-script.js` | DevTools global hook | `rdt` uses the hook for renderer/root discovery, but does not speak the official frontend protocol. |
 | `state.roots` + `rootId` | `src/runtime-script.js` | Fiber roots / renderer root registry | `rdt` assigns CLI-friendly `root-*` ids per discovered root. |
 | `collectTree()` snapshot | `src/runtime-script.js` | Inspected tree payload | `rdt` serializes a tree for CLI use instead of streaming updates to a frontend. |
+| `tree stats` | `src/runtime-script.js` + `src/cli.js` | No direct public equivalent | Snapshot summary view for large trees; returns metadata without the full node dump. |
 | `snapshotId` | `src/runtime-script.js` | No direct public equivalent | Intentional divergence. This is a CLI-specific stability layer for follow-up commands. |
 | `node id` like `n68` | `src/runtime-script.js` | Element/fiber identity in inspector payloads | Snapshot-scoped only. Not intended to be globally stable across commits. |
 | `node inspect` payload | `src/runtime-script.js` | Inspected element details | Closest conceptual match to DevTools inspected element data. |
 | `ownerStack` | `src/runtime-script.js` | Owner chain / component stack | Serialized as lightweight `{id, displayName}` records. |
 | `hooks` | `src/runtime-script.js` | Hook inspection data | Derived from `memoizedState`; intentionally simpler than full DevTools hook typing. |
 | `context` | `src/runtime-script.js` | Context dependencies / inspected context values | Derived from fiber dependencies, serialized for CLI use. |
-| `source reveal` | `src/server.js` + `src/runtime-script.js` | View source / inspect source location | Depends on `_debugSource`; may legitimately return `null`. |
+| `source reveal` | `src/server.js` + `src/runtime-script.js` | View source / inspect source location | Raw mode may legitimately return `null`; structured mode reports source availability explicitly. |
 | `profiler` summary/export | `src/runtime-script.js` + `src/cli.js` | Profiler commit data | `rdt` keeps commit-oriented summaries and NDJSON export rather than full DevTools frontend state. |
 | `profiler compare` | `src/cli.js` | No direct public equivalent | CLI-side comparison of stored profiler artifacts or exported NDJSON files. |
 | `interact click|type|press|wait` | `src/server.js` | No direct public equivalent | Playwright-backed deterministic interaction helpers for agent workflows. |
+| `interact click --delivery` | `src/server.js` | No direct public equivalent | CLI-specific interaction contract for choosing Playwright pointer input vs DOM dispatch. |
 | `session doctor` | `src/server.js` + `src/runtime-script.js` | No direct public equivalent | CLI-specific preflight that reports trust boundaries, runtime readiness, Playwright resolution diagnostics, and helper import targets. |
 
 ## Intentional Divergences
@@ -72,6 +74,8 @@ It is a maintenance aid for agents and contributors. It is not a commitment to r
 - `hooks`: serialized hook state view, intentionally simplified.
 - `context`: serialized current context dependency values.
 - `source`: `_debugSource` projection when available; `null` is expected in many builds.
+- `source reveal --structured`: wraps source lookup in `{ status, available, mode, reason, source }`.
+- `mode`: source-capability mode, not engine mode. Current values describe whether `_debugSource` is available, removed by newer React runtimes, or stripped by the build.
 - `dom`: first host element descendant summary used for CLI-oriented highlight/reveal behavior.
 
 ### Profiler payload

docs/workflows.md

@@ -23,7 +23,8 @@ npm install -g react-devtool-cli
 ```bash
 rdt session open --url http://localhost:3000 --browser chromium --engine auto --session demo
 rdt tree get --session demo
-rdt node search App --session demo --snapshot <snapshotId>
+rdt tree stats --session demo --top 5
+rdt node search App --session demo --snapshot <snapshotId> --structured
 rdt node inspect <nodeId> --session demo --snapshot <snapshotId>
 ```
 
@@ -40,11 +41,16 @@ Recommended flow:
 
 ```bash
 rdt tree get --session demo
-rdt node search App --session demo --snapshot <snapshotId>
+rdt tree stats --session demo --top 5
+rdt node search App --session demo --snapshot <snapshotId> --structured
 rdt node highlight <nodeId> --session demo --snapshot <snapshotId>
-rdt source reveal <nodeId> --session demo --snapshot <snapshotId>
+rdt source reveal <nodeId> --session demo --snapshot <snapshotId> --structured
 ```
 
+- `tree stats` is the lightweight summary path when `tree get --format json` is too heavy.
+- `node search --structured` keeps the default array-returning behavior opt-in while adding `matchCount` and `runtimeWarnings`.
+- `source reveal --structured` returns availability metadata instead of only raw `null`.
+
 Recovery flow:
 
 ```bash
@@ -56,6 +62,7 @@ rdt node search App --session demo --snapshot <newSnapshotId>
 ## Run `doctor` before deeper investigation
 
 ```bash
+rdt doctor --session demo
 rdt session doctor --session demo
 ```
 
@@ -71,11 +78,14 @@ Use it to confirm:
 Built-in interactions keep the investigation inside the same session instead of forcing separate helper scripts.
 
 ```bash
-rdt interact click --session demo --selector 'button.save'
+rdt interact click --session demo --selector 'button.save' --delivery auto
 rdt interact type --session demo --selector 'input[name="query"]' --text hello
 rdt interact wait --session demo --ms 500
 ```
 
+- `interact click --delivery auto` uses Playwright pointer input by default.
+- When the profiler is active, `auto` may fall back to DOM dispatch and reports the applied delivery in the response payload.
+
 After interaction, verify the app settled by collecting a fresh tree or reading profiler output instead of assuming the UI state changed correctly.
 
 ## Profile a real update

package.json

@@ -1,6 +1,6 @@
 {
   "name": "react-devtool-cli",
-  "version": "0.1.35-rc.1",
+  "version": "0.1.35-rc.2",
   "description": "Agent-first CLI for React component tree inspection, snapshot-aware node debugging, and profiler analysis through a Playwright-managed browser session.",
   "license": "MIT",
   "type": "module",

src/cli.js

@@ -92,6 +92,10 @@ function collectSnapshotPayload(options) {
   return options.snapshot ? { snapshotId: String(options.snapshot) } : {};
 }
 
+function normalizeStructuredFlag(options) {
+  return Boolean(options.structured);
+}
+
 function resolveCommitId(positionals, options, message) {
   const commitId = positionals[0] ? String(positionals[0]) : (options.commit ? String(options.commit) : null);
   ensure(commitId, message, { code: "missing-commit-id" });
@@ -278,9 +282,14 @@ async function handleSessionCommand(command, options) {
 }
 
 async function handleTreeCommand(command, options) {
-  ensure(command === "get", "Only `rdt tree get` is supported.", { code: "unsupported-command" });
+  ensure(command === "get" || command === "stats", "Only `rdt tree get` and `rdt tree stats` are supported.", {
+    code: "unsupported-command",
+  });
   ensure(options.session, "Missing required option --session", { code: "missing-session" });
-  const response = await requestSession(options.session, "tree.get");
+  const action = command === "stats" ? "tree.stats" : "tree.get";
+  const response = await requestSession(options.session, action, {
+    top: options.top ? Number(options.top) : undefined,
+  });
   writeStdout(response.result, resolveFormat(options));
 }
 
@@ -304,6 +313,7 @@ async function handleNodeCommand(command, positionals, options) {
     ensure(query, "Missing query for `rdt node search`.", { code: "missing-query" });
     const response = await requestSession(options.session, "node.search", {
       query,
+      structured: normalizeStructuredFlag(options),
       ...collectSnapshotPayload(options),
     });
     writeStdout(response.result, resolveFormat(options));
@@ -339,6 +349,7 @@ async function handleInteractCommand(command, options) {
     ensure(options.selector, "Missing required option --selector for `rdt interact click`.", { code: "missing-selector" });
     const response = await requestSession(options.session, "interact.click", {
       selector: String(options.selector),
+      delivery: options.delivery ? String(options.delivery) : undefined,
       timeoutMs: options.timeoutMs ?? undefined,
     });
     writeStdout(response.result, resolveFormat(options));
@@ -732,11 +743,30 @@ async function handleSourceCommand(command, positionals, options) {
   ensure(nodeId, "Missing node id for `rdt source reveal`.", { code: "missing-node-id" });
   const response = await requestSession(options.session, "source.reveal", {
     nodeId,
+    structured: normalizeStructuredFlag(options),
     ...collectSnapshotPayload(options),
   });
   writeStdout(response.result, resolveFormat(options));
 }
 
+export function normalizeCliPositionals(positionals) {
+  const [resource, command, ...rest] = positionals;
+
+  if (resource === "doctor") {
+    return {
+      resource: "session",
+      command: "doctor",
+      rest: command ? [command, ...rest] : [],
+    };
+  }
+
+  return {
+    resource,
+    command,
+    rest,
+  };
+}
+
 function printHelp() {
   process.stdout.write(`react-devtool-cli
 
@@ -746,21 +776,23 @@ Recommended flow:
   3. Use rdt session attach only for Chromium CDP compatibility
   4. Use rdt tree/node/profiler commands for structured output
   5. For agent workflows, capture snapshotId from tree get and pass it to later node/source commands
-  6. Use rdt doctor before profiling if helper scripts or Playwright resolution look suspicious
+  6. Use rdt doctor (alias: rdt session doctor) before profiling if helper scripts or Playwright resolution look suspicious
 
 Usage:
+  rdt doctor --session <name> [--format json|yaml|pretty]
   rdt session open --url <url> [--browser chromium|firefox|webkit] [--engine auto|custom|devtools] [--channel <name>] [--device <name>] [--storage-state <path>] [--user-data-dir <path>] [--timeout <ms>] [--headless=false] [--session <name>]
   rdt session connect --ws-endpoint <url> [--browser chromium|firefox|webkit] [--engine auto|custom|devtools] [--target-url <substring>] [--timeout <ms>] [--session <name>]
   rdt session attach --cdp-url <url> [--engine auto|custom|devtools] [--target-url <substring>] [--timeout <ms>] [--session <name>]
   rdt session status --session <name>
   rdt session doctor --session <name> [--format json|yaml|pretty]
   rdt session close --session <name>
   rdt tree get --session <name> [--format json|yaml|pretty]
+  rdt tree stats --session <name> [--top <n>] [--format json|yaml|pretty]
   rdt node inspect <id> --session <name> [--snapshot <id>] [--commit <id>]
-  rdt node search <query> --session <name> [--snapshot <id>]
+  rdt node search <query> --session <name> [--snapshot <id>] [--structured]
   rdt node highlight <id> --session <name> [--snapshot <id>]
   rdt node pick --session <name> [--timeout-ms 30000]
-  rdt interact click --session <name> --selector <css> [--timeout-ms <ms>]
+  rdt interact click --session <name> --selector <css> [--delivery auto|playwright|dom] [--timeout-ms <ms>]
   rdt interact type --session <name> --selector <css> --text <value> [--timeout-ms <ms>]
   rdt interact press --session <name> --key <name> [--selector <css>] [--timeout-ms <ms>]
   rdt interact wait --session <name> --ms <n>
@@ -773,7 +805,7 @@ Usage:
   rdt profiler flamegraph <id> --session <name> [--format json|pretty]
   rdt profiler compare --session <name> --left <profileId|file> --right <profileId|file> [--format json|yaml|pretty]
   rdt profiler export --session <name> [--output file.jsonl] [--compress]
-  rdt source reveal <id> --session <name> [--snapshot <id>]
+  rdt source reveal <id> --session <name> [--snapshot <id>] [--structured]
 
 Snapshot behavior:
   - tree get returns snapshotId
@@ -795,7 +827,7 @@ export async function runCli(argv) {
     return;
   }
 
-  const [resource, command, ...rest] = positionals;
+  const { resource, command, rest } = normalizeCliPositionals(positionals);
 
   try {
     switch (resource) {