Enforce a truthful local-first AI/privacy contract
Template tests / tests (pull_request) Failing after 34s

Phase 1 of the improvement plan (PR 3 of the sequence). The docs claimed
"fully offline"/"never talks to the network," but text-intel makes HTTP
requests to a configurable Ollama host that could be remote, with no timeout,
no cancellation, and no size limit; the Windows hook logged raw keystrokes
into capture metadata that could then be sent to that host.

Privacy — raw keystroke capture:
- New capture.captureTypedText setting, default false. With it off, printable
  characters are never buffered in JS and the Windows keyboard hook never even
  emits them across the process boundary (the flag is threaded into the C#).
  Shortcut/navigation detection (Ctrl+T, Enter, …) is unaffected.

AI network hardening (app/text-intel.js):
- Every Ollama call goes through fetchJson with an AbortController deadline
  (ai.timeoutMs, default 60s): a dead endpoint fails fast instead of leaving
  UI actions pending forever.
- Cancellation: in-flight requests are tracked and cancelInflight(guideId)
  aborts them; new ai:cancel IPC + api.ai.cancel are called when the editor
  closes, and shutdown cancels everything.
- Bounded concurrency (2) for AI network work.
- Screenshots are only attached when allowed (ai.attachScreenshots), the model
  is vision-capable, and the image is within ai.maxImageBytes — no more
  unbounded base64-expanded 4K bodies.

Local-first host policy (core/text-intel.js):
- New isLoopbackHost + validateOllamaHost. By default only a loopback Ollama
  endpoint is contacted; a remote host is refused with a clear message unless
  ai.allowRemoteHost is explicitly enabled. Blocked hosts are never contacted.

Honest documentation:
- README, package.json, and the welcome screen drop "fully offline"/"never
  talks to the network"/"Electron is the only dependency" for an accurate
  local-first contract that discloses the optional AI path and the bundled
  Tesseract OCR dependency.
- New docs/PRIVACY.md details exactly what is collected locally and the one
  outbound (opt-in, loopback-by-default) AI feature.

Tests: loopback/remote host matrix, remote-blocked-without-opt-in (and never
contacted), remote-allowed-with-opt-in, request timeout, explicit cancel vs
timeout, typed-text off-by-default vs opted-in, shortcut detection still works,
and a source guard that the C# CHAR emission stays behind the opt-in. 224 unit
tests pass; startup smoke and workflow E2E pass.

Co-Authored-By: Claude Fable 5 <[email protected]>
This commit is contained in:
2026-07-03 11:36:01 -07:00
co-authored by Claude Fable 5
parent 6ffef69705
commit ccbb9b03dc
12 changed files with 533 additions and 33 deletions
+15 -2
View File
@@ -1025,6 +1025,11 @@ class CaptureService {
// by other processes and a polling loop can miss short clicks under
// load; WH_MOUSE_LL gives us one event for each button-down, with the
// hook-time cursor position and timestamp.
//
// Raw typed-text capture is a keylogging surface, so the hook only
// emits printable CHAR events when the user explicitly opted in; by
// default the characters never even cross the process boundary.
const captureTypedText = this.settings.get('capture.captureTypedText') ? 'true' : 'false';
const ps = `
$ErrorActionPreference = 'Stop'
Add-Type -TypeDefinition @'
@@ -1054,6 +1059,7 @@ public static class SFHook {
private const uint PROCESS_POWER_THROTTLING_EXECUTION_SPEED = 0x1;
private const uint HIGH_PRIORITY_CLASS = 0x00000080;
private static readonly bool CaptureTypedText = ${captureTypedText};
private static IntPtr hook = IntPtr.Zero;
private static IntPtr keyHook = IntPtr.Zero;
private static LowLevelMouseProc proc = MouseHookCallback;
@@ -1306,8 +1312,10 @@ public static class SFHook {
queue.Enqueue("KEY Escape " + unixMs); signal.Set();
} else if (vk == 0x0D) {
queue.Enqueue("KEY Enter " + unixMs); signal.Set();
} else {
// Map to Unicode character using current keyboard layout + shift state.
} else if (CaptureTypedText) {
// Map to Unicode character using current keyboard layout + shift
// state. Only reached when the user opted into typed-text capture;
// otherwise raw characters are never read or emitted.
byte[] ks = new byte[256];
GetKeyboardState(ks);
var sb = new System.Text.StringBuilder(4);
@@ -1790,6 +1798,11 @@ public static class SFHook {
this._keyLastAt = now;
if (type === 'CHAR') {
// Raw typed-text capture is off by default: it can record passwords or
// other secrets. Only buffer printable characters when the user has
// explicitly opted in via capture.captureTypedText. Shortcut/navigation
// keys below are unaffected.
if (!this.settings.get('capture.captureTypedText')) return;
const ch = typeof data === 'number' ? String.fromCharCode(data) : String(data);
this._keyBuffer = (this._keyBuffer + ch).slice(-200);
} else if (type === 'KEY') {