@ashx-j/lunr-dev 0.2.13-dev.8.1 → 0.2.19-dev.10.1

package/docs/compaction.md

@@ -34,7 +34,9 @@ contextTokens > contextWindow - reserveTokens
 
 By default, `reserveTokens` is 16384 tokens (configurable in `~/.lunr/agent/settings.json` or `<project-dir>/.lunr/settings.json`). This leaves room for the LLM's response.
 
-You can also trigger manually with `/compact [instructions]`, where optional instructions focus the summary.
+For Codex OAuth models, lunR checks the threshold after each complete batch of tool results. If the batch crosses the threshold, lunR compacts before sending the next provider request.
+
+You can also trigger manually with `/compact [instructions]`, where optional instructions focus the summary. Running `/compact` again at an unchanged compaction boundary is a no-op.
 
 ### How It Works
 
@@ -114,7 +116,7 @@ Valid cut points are:
 - BashExecution messages
 - Custom messages (custom_message, branch_summary)
 
-Never cut at tool results (they must stay with their tool call).
+Never cut at tool results (they must stay with their tool call). If the recent-token budget lands on a tool result, lunR keeps the assistant message containing its tool call.
 
 ### CompactionEntry Structure
 
@@ -288,7 +290,7 @@ pi.on("session_before_compact", async (event, ctx) => {
 
   // branchEntries - all entries on current branch (for custom state)
   // reason - "manual" (/compact), "threshold", or "overflow"
-  // willRetry - whether the aborted turn is retried after compaction (overflow recovery)
+  // willRetry - whether the active run sends another provider request after compaction
   // signal - AbortSignal (pass to LLM calls)
 
   // Cancel:

package/docs/computer-use.md

@@ -0,0 +1,277 @@
+# Native computer use
+
+This implementation is available through the experimental `@ashx-j/lunr-dev`
+channel, based on stable lunR 0.2.19 plus PR #77. Install with
+`npm i -g @ashx-j/lunr-dev`, then run `lunr-dev`. Stable `@ashx-j/lunr` is unchanged.
+This is not production support certification. Windows
+MCP metadata interoperability and local runtime/lifecycle tests pass. The macOS
+launch path is implemented but has not run on macOS hardware. Neither platform
+has passed live desktop acceptance. Production publication remains blocked.
+
+## Local setup
+
+Run `/computer setup` in the local lunR terminal. It verifies and installs the
+bundled runtime without downloading code, capturing the screen, or operating an
+application. Setup is user-invoked and unavailable from gateway or headless
+sessions.
+
+On Windows, keep the desktop logged in and unlocked. On macOS, setup prints the
+exact installed `CuaDriver.app` path. Add that app under System Settings > Privacy
+& Security to both Accessibility and Screen & System Audio Recording, named
+Screen Recording on older macOS versions. Enable both grants, then restart lunR.
+These OS grants require the user; Auto mode cannot supply them. Setup never
+selects or changes an existing `/Applications/CuaDriver.app` installation.
+
+Complete setup before assigning GUI tasks through cron or the gateway. A setup
+success verifies the runtime files, not permission grants or desktop behavior.
+
+## Tools and permissions
+
+Computer use and Allow foreground control default to on. Windows x64/arm64 and
+Apple Silicon macOS expose `computer_load` before the first model request.
+The runtime loads only when a desktop operation needs it. Linux and Intel macOS
+have no default discovery tool.
+
+Loading activates app/window discovery, accessibility and image observation,
+single/right/double clicks, complete drags, scrolling, keys and modifier
+shortcuts, Unicode text, app launch, window management, and workflow release.
+Use these tools on the computer running lunR. Gateway sessions control the
+gateway host, not the Telegram or Discord client's machine.
+
+Prefer accessibility and background window input. Request window foreground
+input only after a background failure and a fresh observation. Primary-desktop
+input has no background route and requires `desktop=true` and `foreground=true`.
+Foreground off blocks desktop input, explicit foreground window input, app
+launches, and every window-management operation. Background delivery cannot
+prevent an application from changing its own focus.
+
+`computer_window` requires an exact window target and has no per-call foreground
+or desktop option. It moves/resizes with `action=frame`, or activates the exact
+window with `action=focus`. Minimize/restore invokes the supplied, freshly
+observed accessibility window-control element. The pinned driver has no portable
+minimize/restore RPC. If the control or minimized window cannot be observed, use
+freshly observed desktop controls rather than guessing a shortcut or window id.
+Text and keys can target an accessibility element, window-image coordinates,
+or the observed focused field. Desktop keyboard input targets the observed
+focused field; changing focus needs a separate grounded click.
+
+Manual asks before every computer call except `computer_end`, which only releases
+the workflow and never needs approval. Observation still requires approval.
+Plan permits relevant observation and workflow release but blocks mutation. Auto needs no
+lunR per-call prompt for requested, implied, or necessary GUI work. In Yolo,
+ask first when an otherwise non-GUI task newly requires GUI. This intent rule
+is agent guidance, not a natural-language authorization classifier.
+
+Native tools are excluded from children and also refuse child execution.
+Shell-capable children are not OS-sandboxed. No generic MCP registration or
+raw driver policy, update, recording, or administrative tool is exposed.
+
+## Grounding and outcomes
+
+Each observation invalidates the previous token before contacting the driver.
+A token identifies one exact window or the primary desktop, expires after thirty
+seconds, and permits one action. Accessibility input includes the driver's
+snapshot id. Structured refusals count as refusals even without MCP `isError`.
+Partial and unverifiable outcomes remain visible; transport success is not
+proof of an application change.
+
+Window PNGs are capped by the driver at 1024 pixels on their long edge.
+Window coordinates use that returned image, without another resize. Both pinned
+platform implementations retain the capture resize ratio for later pixel input.
+The adapter checks PNG dimensions and refuses inconsistent pixel grounding. Desktop
+images are resized to at most 1024 pixels locally. Their token records separate
+x/y ratios back to the driver's native desktop image. Coordinates outside the
+returned image are rejected. Pixel and desktop operations require an
+image-capable current model; accessibility-only observation remains available.
+
+Observe after every action. A timeout or cancelled call may already have
+changed the application. Never retry input blindly. Cancelling the MCP request
+does not retract input already delivered to the OS. Normal shutdown asks the
+owned runtime to drain and exit; forced process termination cannot prove that
+all held input was released. Live gesture cancellation remains an acceptance
+gate.
+
+## Ownership and lifecycle
+
+An OS-account lease holds the desktop across observe/action calls. Competing
+workflows get busy rather than queueing; calls within one workflow serialize.
+The lease uses the OS account home, not the selected lunR settings profile.
+One short cross-process acquisition lock protects creating, updating, and deleting
+the owner record. Updates replace the file atomically. Runtime installation uses
+a separate lock and never replaces an existing runtime. A live owner is never displaced because its heartbeat is late.
+After MCP initialization and before any driver tool call, the adapter records
+the transport PID and, on macOS, the daemon PID in that lease. Missing PIDs or
+failure to record ownership block tool dispatch. Driver-internal helper lifetime
+remains part of the unverified native shutdown acceptance gates.
+A crash during initialization can leave an unrecorded, idle runtime, but it has
+received no desktop action. Recovery requires both the owner and its recorded
+runtimes to have exited.
+PID reuse can conservatively leave the desktop busy rather than kill another
+process. A malformed owner record also fails closed.
+
+`computer_end`, cancellation, agent end, settings changes, and session shutdown
+close the owned runtime before releasing its lease. Queued operations cannot
+survive cancellation. Settings disable removes active tools immediately;
+session replacement discards loaded tools and grounding. The first-request
+hook also reapplies the loaded roster after extension registration refreshes it.
+If shutdown cannot be confirmed, the lease stays held. Confirm that the owned
+runtime has stopped before restarting the owning lunR process.
+
+Fresh gateway cron sessions get explicit permission contexts from the configured
+default mode. They inherit neither gateway approvals nor its interactive
+approval handler. Missing approval reports blocked. Session shutdown runs before
+context disposal, including partial extension-bind failure. TUI cron continues
+to use its live session mode.
+
+## Runtime and packaging
+
+The development pin is [CuaDriver 0.28.1](https://github.com/trycua/cua/releases/tag/cua-driver-rs-v0.28.1),
+source `d8028a7943087ee258dc1b4d19dc12a7cd27669c`. This exact prerelease and its
+three archive hashes are approved for development-channel distribution only.
+`scripts/computer-use-release.json` is the authoritative archive inventory.
+The upstream MIT license is included under `native/computer-use`.
+
+The runtime never downloads an executable. It resolves the host's optional
+payload package, or the matching archive beside a standalone binary. It verifies that archive,
+extracts a fresh reference tree, and checks every cached file and helper against
+that tree before execution. Redirected install paths and escaping symlinks are
+rejected. Concurrent installers do not overwrite each other. A modified or
+older cache is refused rather than overwritten. After confirming no runtime is
+active, remove the inactive lunR runtime directory to reinstall it.
+
+Windows uses the system `tar.exe`, not a `PATH`-selected extractor, and launches
+`mcp --direct --embedded --no-overlay`. A bounded, hidden system PowerShell
+helper checks WTS session activity and the input desktop name before admission.
+It refuses session zero, disconnected sessions, and a non-Default or inaccessible
+input desktop. It reads metadata only and never switches desktops. The isolated
+runtime profile avoids the user's Cua config. Both platforms remove inherited `CUA_*` settings and explicitly
+set standard permission mode, disable unrestricted mode, and set
+`CUA_DRIVER_RS_TELEMETRY_ENABLED=false` and `CUA_DRIVER_RS_UPDATE_CHECK=false`.
+
+macOS preserves the signed `CuaDriver.app` at the stable lunR-owned path
+`~/.lunr/desktop/runtime-darwin-arm64/CuaDriver.app` and verifies its signature
+with `codesign --verify --deep --strict`. It never re-signs the bundle.
+LaunchServices selects that exact app path with `open -n -g -W -a`, not its
+shared bundle id or the user's `/Applications` installation. Explicit `--env`
+arguments pass telemetry/update/permission settings and a private HOME to the
+daemon. The private HOME also isolates the upstream default PID/config files.
+A private Unix socket carries the MCP proxy connection. A private stdin FIFO
+and the pinned `--parent-liveness-stdio` option tie daemon lifetime to lunR.
+Shutdown uses the private endpoint's PID-bound request and waits for the owned
+app to exit. Startup and shutdown have time limits. These construction and
+cleanup paths have local tests, not macOS execution evidence.
+
+Release staging creates `@ashx-j/lunr-computer-win32-x64`,
+`@ashx-j/lunr-computer-win32-arm64`, and
+`@ashx-j/lunr-computer-darwin-arm64`. Each package contains one unchanged archive,
+the MIT license, and a manifest with `os` and `cpu`. There are no install scripts.
+The payload packages follow the CLI release version, independently of the upstream
+runtime version. The staged CLI declares exact optional versions. npm installs
+only the matching payload; `--ignore-scripts` works. Missing optional dependencies
+leave the CLI usable and produce a reinstall instruction when computer use needs
+the runtime. Unsupported hosts have no native discovery tool.
+
+The CLI tarball is about 10.7 MB, without runtime archives. Packed payload sizes
+are about 28.8 MB for Windows x64, 27.2 MB for Windows arm64, and 69.8 MB for
+Apple Silicon. These are decimal sizes, not extracted runtime sizes.
+
+Workspace manifests and locks keep unpublished native dependencies out of
+ordinary developer installs. `scripts/publish.mjs` injects them into the staged
+CLI shrinkwrap and standalone installer lock. Public lock rewriting updates both
+package names and tarball filenames. Publication validates all seven packages,
+then publishes payloads before the CLI. Dev staging uses exact
+`0.2.19-dev.<run>.<attempt>` versions throughout package manifests, shrinkwrap,
+and installer locks. It publishes `@ashx-j/lunr-dev` with its own `latest` tag;
+shared libraries and runtime payloads use `dev` tags. Stable tags stay unchanged.
+Production publication still requires separate approval and an unpublished
+stable CLI version.
+
+`scripts/build-binaries.sh` and `copy-binary-assets` copy only the selected opaque
+archive into `native/computer-use` beside the executable. They preserve the macOS
+archive without extracting or signing it. Routine builds stay offline. Release
+preparation may fetch only the pinned official archives, then verifies their
+sizes and hashes. Stable npm and GitHub binary publication stop on the current
+development-only approval. The `dev/tui` workflow may publish the explicitly
+requested npm dev channel.
+
+An installed payload package must match the CLI version. A CLI upgrade that keeps
+the same runtime bytes can reuse the verified cache. A changed or older cached
+runtime fails closed rather than replacing a possibly running app. After all
+owned processes stop, removing the inactive lunR runtime directory allows a fresh
+extraction at the same stable macOS path.
+
+## Verification and remaining gates
+
+Local verification includes actual Windows initialize/tools/list and a
+metadata-only `health_report`, with capture and accessibility checks excluded.
+The binary reported 0.28.1 and 57 tools. Health results include human-readable
+text and `structuredContent`. Its `session_active` check reports MCP activity,
+not an unlocked or input-ready desktop. A separate real Windows metadata probe
+confirmed an active session and Default input desktop without screen, AX, or
+input calls. Snapshot identity and operation mapping
+were checked against the exact released tool schemas and source; no live
+snapshot was taken.
+
+Focused tests cover policy, result interpretation, image coordinate mapping,
+new GUI operation mapping, interrupted initialization, partial shutdown,
+settings/session tool rosters, real separate-process contention and death,
+orphan-runtime ownership, cache tampering, redirected paths, concurrent
+extraction, and the fresh gateway cron factory's approval/disposal lifecycle.
+The dev review passes 167 focused Vitest tests across 14 files and
+eight archive/package tests. All five offline production builds, including the
+coding-agent Node bundle, pass. Touched native TypeScript lint and the repository's
+relative-import, workflow-publication, browser-smoke, and diff checks pass.
+
+A loopback fixture registry served all seven public-name tarballs without changing
+their bytes. A fresh Windows install used `--ignore-scripts`, fetched only the host
+payload, then passed archive validation and first-paint/first-request fixtures
+after relocation. Separate npm installs selected the correct package for each
+supported OS/CPU combination and no package for Linux. An omitted-optional install
+returned the expected missing-payload error. The shipped installer lock passed
+`npm ci --ignore-scripts`. Standalone asset tests copied and rechecked all three
+opaque archives. No installed app or desktop operation ran in these packaging
+tests. Cross-platform npm selection on Windows is not macOS runtime acceptance.
+
+The first-request fixture strips inherited subagent environment state before
+checking the default tool inventory. The supported-host inventory includes only
+`computer_load`; detailed computer tools remain lazy. The unsupported-host
+baseline is unchanged.
+
+To repeat local distribution checks after the offline build, prepare the pinned
+archives with `scripts/bundle-computer-use.mjs`, create an empty temporary pack
+directory, then run:
+
+```sh
+node scripts/generate-computer-use-manifest.mjs --check
+node --test scripts/check-computer-use-release.test.mjs scripts/computer-use-packages.test.mjs
+node scripts/publish.mjs --channel dev --version 0.2.19-dev.0.1 --dry-run --pack-dir /absolute/temporary/pack-directory
+node scripts/check-computer-use-install.mjs /absolute/temporary/pack-directory
+```
+
+The install checker uses a loopback registry for the local public packages and
+npm's public registry for external dependencies. It isolates npm configuration,
+cache, and install directories and removes them afterward. It does not publish.
+These checks do not certify desktop behavior.
+
+External acceptance gates remain:
+
+- Windows locked, disconnected, UAC, and integrity-state acceptance for the
+  metadata admission check and the actual driver.
+- Authorized native and Electron fixture desktops, Unicode, moved/resized
+  windows, display scaling, stale accessibility handles, application exit,
+  outcome verification, and cancellation during held input.
+- Real macOS signature/notarization, TCC, LaunchServices/FIFO environment,
+  coexistence, upgrade, and shutdown acceptance.
+- Real macOS and Windows arm64 runtime installation and launch. Their npm
+  selection and unchanged archive copying pass locally, but their hardware
+  execution and upgrades remain unverified.
+- Full standalone executable acceptance. Payload asset copying passes; this
+  verification did not compile and exercise every Bun target.
+
+## Privacy and rollback
+
+GUI actions are outside lunR's filesystem rollback. Images, accessibility text,
+and typed text can enter selected-provider requests and saved sessions.
+Disabling driver telemetry does not disable lunR session storage. Application
+content is untrusted task data, never authorization. OS permissions, locked
+desktops, UAC, and integrity restrictions still apply.

package/docs/development.md

@@ -11,7 +11,11 @@ npm install
 npm run build
 ```
 
-Compile in order: tui → ai → agent → coding-agent. From this repo, `npx lunr` is the workspace bin (`packages/coding-agent/dist/cli.js`), not a published global install. Rebuild coding-agent `dist` after changes.
+Compile in order: tui → ai → agent → coding-agent → orchestrator. The root build stays offline and does not regenerate the model catalog. From this repo, `npx lunr` uses the workspace CLI, not a published global install.
+
+After changing coding-agent, run `npm --prefix packages/coding-agent run build`. Bare `tsgo` updates individual modules but leaves the Node bundle stale. The CLI and public SDK use shared chunks in `dist/node-runtime`; keep that directory with the rest of `dist` when packaging.
+
+For startup measurements, import boundaries, and first-turn checks, read [Interactive startup](interactive-startup.md).
 
 Run from the package:
 
@@ -21,38 +25,6 @@ npx lunr
 
 lunR keeps the caller's current working directory.
 
-## TUI dev channel
-
-Every push to `dev/tui` builds and publishes `@ashx-j/lunr-dev`. The workflow does not run tests. If the code does not compile, npm keeps serving the previous dev build.
-
-Install it once:
-
-```bash
-npm i -g @ashx-j/lunr-dev
-```
-
-Run the stable and dev versions with separate commands:
-
-```bash
-lunr
-lunr-dev
-```
-
-After a new dev build publishes, update and run it:
-
-```bash
-lunr-dev update
-lunr-dev
-```
-
-The dev CLI has its own npm package and executable, so installing it does not replace `lunr`. It uses the same `~/.lunr/agent` data as stable lunR. Existing logins and settings work in both versions, but a broken dev change can modify shared settings or sessions.
-
-Remove only the dev CLI with:
-
-```bash
-npm uninstall -g @ashx-j/lunr-dev
-```
-
 ## Forking / Rebranding
 
 Configure via `package.json`:

package/docs/features.md

@@ -6,7 +6,7 @@ Sample code under `examples/extensions/plan-mode`, `examples/extensions/todo.ts`
 
 ## Permissions and plan mode
 
-Permission modes: `manual | yolo | plan | auto`. Shift+Tab (`app.mode.cycle`) cycles that order. Thinking cycle is unbound; use `/thinking`, `/effort`, or `/reasoning`.
+Permission modes: `manual | yolo | plan | auto`. Shift+Tab (`app.mode.cycle`) cycles that order. Ctrl+T (`app.thinking.cycle`) cycles thinking levels for the selected model.
 
 - **manual** — approve every tool
 - **yolo** — auto-approve ordinary tools; large subagent launches still request confirmation
@@ -23,7 +23,6 @@ Advertised subagents always start **fresh** (no forked parent context). Default
 
 A launch of 3+ parallel children in one `tasks`/`chain.parallel` call, or 3+ same-turn SINGLE `subagent` calls, receives one aggregate confirmation in **manual and yolo**. Sequential work stays `chain`. Auto bypasses this confirmation, and it can be disabled independently in `/settings`.
 
-
 `/goal` sets a session goal and **forces session auto** permission mode.
 
 ## Todos, memory, and global instructions
@@ -79,7 +78,7 @@ Without runnable adapters (enabled platform + resolvable token), `lunr gateway`
 
 ## Thinking, usage, streaming, UI
 
-- `/thinking`, `/effort`, and `/reasoning` are full-parity aliases. `/thinking` only offers levels the session model supports (`getSupportedThinkingLevels`). `xhigh` and `max` are opt-in.
+- `/thinking`, `/effort`, and `/reasoning` are full-parity aliases. `/off`, `/minimal`, `/low`, `/medium`, `/high`, `/xhigh`, and `/max` set a level when the current model supports it. `/thinking` completions follow the session model. `xhigh` and `max` are opt-in. `/thinking hide|show|toggle` still hides thinking blocks.
 - `/usage` is **this-session** context plus every stored-credential subscription plan. The current provider is included for env-only auth. There is no `/token-usage`.
 - `/fast [on|off|status]` controls `service_tier: "fast"` for OpenAI Codex subscriptions only. It persists across new sessions, gateway turns, and subagents. Paid `openai` API models do not use it.
 - Footer plan bar prefers a 5h window and falls back to weekly (`planUsageWindow`). In Customize, Plan usage hides the whole segment while Plan bar hides only the █░ fill and keeps the percent.

package/docs/interactive-startup.md

@@ -0,0 +1,105 @@
+# Interactive startup
+
+lunR paints the normal moon chatbox before loading the agent runtime. It keeps
+that terminal and editor through attachment. Enter holds the editable draft until
+session setup and required extension hooks finish. Escape cancels a pending
+submission. Failed initialization preserves the draft and reports the error.
+
+First paint and first request are separate measurements. An early chatbox does
+not mean the agent can send a request yet.
+
+## First-request loading
+
+The Node build uses split ESM bundles in `dist/node-runtime`. CLI and SDK entry
+points share chunks so extensions see the same SDK classes as the running CLI.
+The package RPC export uses a small wrapper that prefers its bundle and falls
+back to the unbundled entry after a TypeScript-only development build. The
+original `dist` layout remains available for assets and native dependencies.
+The build preserves original module URLs for asset lookup and extension aliases.
+
+Startup still registers tools, commands, permissions, and session hooks before
+opening the prompt barrier. It does not replace tool schemas with placeholders.
+The expensive implementations load when needed:
+
+- Web tools load providers, extraction, and browser curation separately.
+- LSP registers its tools before loading clients and Tree-sitter. Configured
+  autostart runs in the background; first use waits for the required services.
+- MCP registers the proxy and cached direct-tool definitions without loading its
+  connection engine. Configured eager servers, keep-alive, and metadata discovery
+  retain their background behavior.
+- Subagents keep restoration and watchers at startup, but load the executor on
+  first use.
+- Third-party extensions load jiti when needed. Bun's static extension host keeps
+  its embedded modules. Package-management commands and HTML export load their
+  implementations only when invoked.
+
+Session replacement invalidates pending initialization. MCP cancels initialization
+and closes partially connected state. LSP drops stale write diagnostics. Web
+requests and curator callbacks cannot publish into a replacement session.
+
+The Node bundle does not automatically enable Node's compile cache. Reusing that
+cache was slower in the measured bundle. Node's explicit compile-cache environment
+settings remain available; the unbundled launch path keeps its existing policy.
+
+## Validation
+
+Run the offline build, including coding-agent's bundle step, then:
+
+```sh
+node scripts/check-interactive-first-paint.mjs
+node scripts/profile-coding-agent-node.mjs --mode tui --skip-build --runs 10
+```
+
+The check stalls or fails runtime loading and verifies the real frame, editable
+draft, and terminal cleanup. It also blocks optional implementation imports and
+checks a complete first request against the baseline tool-payload hash. Separate
+first-turn fixtures exercise subagent status, MCP status, Tree-sitter parsing,
+and local HTTP extraction. The default CI run checks the workspace build on
+Ubuntu. Passing a CLI path runs the same checks against a relocated installation,
+but CI does not currently pack that installation or run this check on Windows.
+
+The publish script validates every staged compiled file after rewriting package
+names and refuses missing package entry points or leftover `@earendil-works/*`
+references.
+
+The profiler uses a local Faux provider, not a remote model. It records dispatch
+through the normal prompt path, plus first response and optional tool completion.
+It isolates the home directory, settings, temporary files, and workspace by
+default. It reports tool names and a schema hash, not prompt contents.
+
+Use `--cli /path/to/dist/cli.js` to compare another built version. Use a dedicated
+`--agent-dir` with `--warmup 1` for repeated launches. Without `--agent-dir`, each
+run gets a fresh profile, including warmup runs. These are fresh profile and
+compile-cache measurements, not reboot-cold filesystem measurements.
+
+## Measured results
+
+Windows, Node 24.15.0, offline, empty workspace, builtin extensions, 2026-09-08.
+Baseline is `b709d12` with benchmark instrumentation only. Measurements preceded
+the rebase onto v0.2.16; build and first-turn checks were repeated afterward.
+Ten launches per cell
+alternated baseline and changed builds. Repeated launches reused each variant's
+own profile after one warmup. The last repeated baseline sample ran separately
+after the comparison runner reached its time limit.
+
+| Profile | Baseline request median / p95 | Changed request median / p95 | Median reduction |
+| --- | --- | --- | --- |
+| Fresh | 1988.8 / 2026.4 ms | 968.2 / 1007.0 ms | 51.3% |
+| Repeated | 2597.6 / 2628.6 ms | 970.6 / 993.5 ms | 62.6% |
+
+First-frame medians were 100.7 to 90.0 ms for fresh profiles and 107.2 to 90.8 ms
+for repeated launches. With ten samples, nearest-rank p95 is the maximum.
+
+First-use fixtures on Node 24 added 68 ms for subagent status, 3075 ms for MCP
+status, 143 ms for LSP parsing, and 257 ms for local HTTP extraction after request
+dispatch. These are single observations of complete tool execution, not isolated
+import costs. Later checks measured 126 to 482 ms for MCP status, so the first-use
+observations are not a stable latency bound.
+
+The packaged CLI passed the same checks on Node 22.19.0. A separately instrumented
+Node 22 run took 6 seconds to dispatch, so these medians are not a latency bound.
+Remote provider latency, terminal compositor latency, reboot-cold storage, large
+user configurations, and third-party asynchronous extension hooks were not
+measured. Bun 1.4.2 source launch completed a local request. Standalone compilation
+was blocked on both baseline and changed builds by the missing native canvas
+binary in the isolated, ignore-scripts dependency installation.

package/docs/keybindings.md

@@ -112,8 +112,8 @@ Modifier combinations: `ctrl+shift+x`, `alt+ctrl+x`, `ctrl+shift+alt+x`, `ctrl+1
 | `app.model.cycleForward` | `ctrl+p` | Cycle to next model |
 | `app.model.cycleBackward` | `shift+ctrl+p` | Cycle to previous model |
 | `app.mode.cycle` | `shift+tab` | Cycle permission mode (`manual` → `yolo` → `plan` → `auto`) |
-| `app.thinking.cycle` | *(unbound)* | Cycle thinking level (`/thinking`, `/effort`, `/reasoning`, and `/settings` still set it) |
-| `app.thinking.toggle` | `ctrl+t` | Collapse or expand thinking blocks |
+| `app.thinking.cycle` | `ctrl+t` | Cycle thinking level for the selected model |
+| `app.thinking.toggle` | *(unbound)* | Collapse or expand thinking blocks (`/thinking hide`, `show`, or `toggle`) |
 
 ### Display and Message Queue
 

package/docs/terminal-setup.md

@@ -108,7 +108,7 @@ On Windows, VS Code owns `Ctrl+V` and uses `Alt+V` as the View menu mnemonic. Fo
 }
 ```
 
-A successful paste toasts `Pasted [image_1]` and inserts that chip in the editor. If VS Code still owns the key, `/paste-image` invokes the same operation without depending on any terminal shortcut. Submit keeps the label in the chat card and attaches the image to the model. `Clipboard contains no image` means the command arrived but the OS clipboard had no image.
+A successful paste inserts `[image_1]` in the editor. If VS Code still owns the key, `/paste-image` invokes the same operation without depending on any terminal shortcut. Submit keeps the label in the chat card and attaches the image to the model. `Clipboard contains no image` means the command arrived but the OS clipboard had no image.
 
 ## Windows Terminal
 

package/docs/usage.md

@@ -69,7 +69,6 @@ Built-in commands (from `slash-commands.ts`):
 | `/logout`, `/deauth` | Remove provider authentication |
 | `/new` | Start a new session |
 | `/init` | Generate a starter AGENTS.md for this project |
-| `/swarm` | Orchestrate parallel subagents for a complex task |
 | `/compact` | Manually compact the session context |
 | `/resume`, `/sessions` | Browse and resume sessions |
 | `/reload` | Reload keybindings, extensions, skills, prompts, themes, and context files |
@@ -77,7 +76,7 @@ Built-in commands (from `slash-commands.ts`):
 
 There is no `/changelog` command.
 
-Notable extension-registered commands (always available unless you disable those built-ins): `/thinking`, `/effort`, `/reasoning`, `/cron`, `/goal`, `/mcp`, `/mcp-auth`, `/lsp`, `/lsp-restart`, `/websearch`. See [Built-in features](features.md).
+Notable extension-registered commands (always available unless you disable those built-ins): `/thinking`, `/effort`, `/reasoning`, `/off` through `/max`, `/cron`, `/goal`, `/mcp`, `/mcp-auth`, `/lsp`, `/lsp-restart`, `/websearch`. See [Built-in features](features.md).
 
 ## Message Queue
 

package/examples/extensions/custom-provider-anthropic/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "pi-extension-custom-provider",
-  "version": "0.2.13",
+  "version": "0.2.19",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "pi-extension-custom-provider",
-      "version": "0.2.13",
+      "version": "0.2.19",
       "dependencies": {
         "@anthropic-ai/sdk": "^0.52.0"
       }

package/examples/extensions/custom-provider-anthropic/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "pi-extension-custom-provider-anthropic",
 	"private": true,
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/custom-provider-gitlab-duo/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "pi-extension-custom-provider-gitlab-duo",
 	"private": true,
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/gondolin/package-lock.json

@@ -1,12 +1,12 @@
 {
 	"name": "pi-extension-gondolin",
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "pi-extension-gondolin",
-			"version": "0.2.13",
+			"version": "0.2.19",
 			"dependencies": {
 				"@earendil-works/gondolin": "0.12.0"
 			}

package/examples/extensions/gondolin/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "pi-extension-gondolin",
 	"private": true,
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/hidden-thinking-label.ts

@@ -9,7 +9,7 @@
  *
  * Test:
  *   1. Load this extension
- *   2. Hide thinking blocks with Ctrl+T
+ *   2. Hide thinking blocks with /thinking hide
  *   3. Ask for something that produces reasoning output
  *   4. The collapsed thinking block label will show the custom text
  *

package/examples/extensions/sandbox/package-lock.json

@@ -1,12 +1,12 @@
 {
 	"name": "pi-extension-sandbox",
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "pi-extension-sandbox",
-			"version": "0.2.13",
+			"version": "0.2.19",
 			"dependencies": {
 				"@anthropic-ai/sandbox-runtime": "^0.0.26"
 			}

package/examples/extensions/sandbox/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "pi-extension-sandbox",
 	"private": true,
-	"version": "0.2.13",
+	"version": "0.2.19",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/with-deps/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "pi-extension-with-deps",
-  "version": "0.2.13",
+  "version": "0.2.19",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "pi-extension-with-deps",
-      "version": "0.2.13",
+      "version": "0.2.19",
       "dependencies": {
         "ms": "^2.1.3"
       },