docs(01): record Phase 1 planning complete (7 plans, 7 waves)

After gsd-plan-phase 1: 7 plans across 7 waves. All gates pass:
- Plan-checker (sonnet) VERIFICATION PASSED on iteration 1
- Decision coverage gate (gsd-sdk): 19/19 decisions covered
- Requirements coverage: REQ-video-ring-buffer in all plans
- Security threat model: T-1-01/02/04 mitigated; T-1-03 accepted residual

Known non-blocking gaps:
- gsd-sdk roadmap.annotate-dependencies failed (t.trim is not a function);
  ROADMAP plan-list annotations skipped. Phase 1 plan-list in ROADMAP.md
  remains accurate; this is a cosmetic nice-to-have for cross-cutting
  constraint visibility.
- 1 plan-checker warning (stale wave prose in Plan 03/04 objectives) was
  fixed during decision-coverage revision.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.planning/STATE.md

@@ -2,14 +2,14 @@
 gsd_state_version: 1.0
 milestone: v2.0.0
 milestone_name: milestone
-status: planning
+status: executing
 stopped_at: Phase 1 context gathered
-last_updated: "2026-05-15T13:40:45.486Z"
+last_updated: "2026-05-15T15:08:45.135Z"
-last_activity: 2026-05-15 — bootstrap from intel synthesis (PROJECT.md,
+last_activity: 2026-05-15 -- Phase 1 planning complete
 progress:
   total_phases: 5
   completed_phases: 0
-  total_plans: 0
+  total_plans: 7
   completed_plans: 0
   percent: 0
 ---
@@ -29,8 +29,8 @@ no server, no password leaks.
 
 Phase: 1 of 5 (Stabilize video pipeline)
 Plan: 0 of TBD in current phase
-Status: Ready to plan
+Status: Ready to execute
-Last activity: 2026-05-15 — bootstrap from intel synthesis (PROJECT.md,
+Last activity: 2026-05-15 -- Phase 1 planning complete
 REQUIREMENTS.md, ROADMAP.md, STATE.md written)
 
 Progress: [░░░░░░░░░░] 0%

.planning/phases/01-stabilize-video-pipeline/01-01-PLAN.md

@@ -12,6 +12,7 @@ files_modified:
   - .planning/intel/constraints.md
   - manifest.json
 autonomous: true
+scope_exception: "doc-text-only edits with no TypeScript compilation; cognitive load substantially lower than code plans of equal task count"
 requirements:
   - REQ-video-ring-buffer
 requirements_addressed:
@@ -24,7 +25,7 @@ must_haves:
     - "ROADMAP.md Phase 1 Success Criterion #2 no longer references tab re-attach"
     - "intel/decisions.md DEC-003 and DEC-010 carry an Amendment block"
     - "intel/constraints.md CON-tab-capture-binding and CON-service-worker-keepalive are RETIRED with a CON-display-capture-binding added"
-    - "manifest.json permissions list contains desktopCapture (not tabCapture) and drops the now-unused alarms entry"
+    - "manifest.json permissions list contains desktopCapture (not tabCapture) and drops the now-unused alarms entry (D-05, D-A6)"
     - "Every code-touching plan (02-07) sees a consistent doc baseline before it runs"
   artifacts:
     - path: ".planning/PROJECT.md"

.planning/phases/01-stabilize-video-pipeline/01-02-PLAN.md

@@ -2,7 +2,7 @@
 phase: 01-stabilize-video-pipeline
 plan: 02
 type: execute
-wave: 0
+wave: 1
 depends_on: ["01"]
 files_modified:
   - package.json

.planning/phases/01-stabilize-video-pipeline/01-03-PLAN.md

@@ -2,7 +2,7 @@
 phase: 01-stabilize-video-pipeline
 plan: 03
 type: tdd
-wave: 1
+wave: 2
 depends_on: ["02"]
 files_modified:
   - src/offscreen/recorder.ts
@@ -10,6 +10,7 @@ files_modified:
   - src/shared/logger.ts
   - src/shared/types.ts
 autonomous: true
+style_divergence_note: "OffscreenLogger uses `...args: unknown[]` (strict-mode hygiene) where existing Logger / ContentLogger use `...args: any[]` (project convention). Refactoring the existing two for consistency is OUT OF SCOPE for Phase 1; they remain on the legacy pattern."
 requirements:
   - REQ-video-ring-buffer
 requirements_addressed:
@@ -17,12 +18,14 @@ requirements_addressed:
 
 must_haves:
   truths:
-    - "`src/offscreen/recorder.ts` exists and exports the symbols Plan 02 tests against: addChunk, trimAged, getBuffer, resetBuffer, assertCodecSupported, VIDEO_BUFFER_DURATION_MS"
+    - "`src/offscreen/recorder.ts` exists at the canonical source path as a real TypeScript module — strict type-check, source maps, IDE support (D-06)"
+    - "`src/offscreen/recorder.ts` exports the symbols Plan 02 tests against: addChunk, trimAged, getBuffer, resetBuffer, assertCodecSupported, VIDEO_BUFFER_DURATION_MS"
     - "Running `npx vitest run tests/offscreen/ring-buffer.test.ts tests/offscreen/codec-check.test.ts` exits 0 with all tests green"
     - "Buffer holds at most: 1 header chunk + every chunk with arrival timestamp newer than now-30_000ms"
     - "Codec strictly bound to `video/webm;codecs=vp9` at 400 000 bps with `MediaRecorder.isTypeSupported` gate; no fallback chain (D-20)"
-    - "`MediaRecorder.start(2000)` is called on session start (timeslice = 2000 ms per SPEC §4.1)"
+    - "Capture is `navigator.mediaDevices.getDisplayMedia()` invoked from inside the offscreen document (D-01); SW-side `chrome.tabCapture.getMediaStreamId` is removed in Plan 05"
-    - "On `MediaStreamTrack.onended`, the buffer is cleared and a `RECORDING_ERROR` of `'user-stopped-sharing'` is emitted to SW"
+    - "Single continuous `MediaRecorder` runs for the whole session with `mediaRecorder.start(2000)` so chunks land on cluster boundaries per SPEC §4.1 (D-09)"
+    - "One-time source picker fires on session start (operator picks screen/window once); on `MediaStreamTrack.onended` the buffer is cleared and a `RECORDING_ERROR` of `'user-stopped-sharing'` is emitted to SW so the popup can re-prompt next interaction (D-03)"
     - "Restart-segments fallback (D-13) is pre-staged as a commented-out skeleton at the bottom of recorder.ts so Plan 07's fallback path doesn't require a re-plan"
     - "`src/offscreen/index.html` exists at the source path and references `./recorder.ts`"
     - "`src/shared/logger.ts` has an `OffscreenLogger` class with `[OS:...]` prefix"
@@ -71,8 +74,10 @@ handshake — that is Plan 04. To keep this plan inside its context budget, the
 port-side code in `recorder.ts` is left as a small import-time placeholder
 that Plan 04 fills in. The Plan-02 port + handshake tests therefore remain
 RED until Plan 04 lands, which is the intended choreography per the
-wave-1 dependency graph (Plans 03 and 04 run in parallel; Plan 02 wrote both
+sequential wave structure (Plan 03 lands in wave 2; Plan 04 follows in wave
-sets of RED tests in advance).
+3 and refactors the bootstrap section of the same file). Plan 02 wrote both
+sets of RED tests in advance so Plan 03 and Plan 04 each have a discrete
+RED→GREEN cycle to flip.
 
 Purpose: REQ-video-ring-buffer's load-bearing logic lives here. The ring
 buffer is a pure function — exactly the TDD sweet spot. The remaining
@@ -154,7 +159,7 @@ export interface VideoChunk {
 | Threat ID | Category | Component | Disposition | Mitigation Plan |
 |-----------|----------|-----------|-------------|-----------------|
 | T-1-01 | Tampering — codec downgrade | `MediaRecorder` constructor | mitigate | `assertCodecSupported()` calls `MediaRecorder.isTypeSupported('video/webm;codecs=vp9')` BEFORE constructing the recorder; if it returns false, throws an Error and emits `RECORDING_ERROR` to SW. No vp8 / h264 / default fallback chain. The strict mode covers `MediaRecorder` itself being absent (the codec-check test mocks both cases). Grep gate: `grep -v '^#' src/offscreen/recorder.ts \| grep -cE 'codecs=(vp8\|h264)'` returns 0 (no fallback codec strings in the module). |
-| T-1-03 | Information Disclosure — captured stream contains other apps' content | `getDisplayMedia` stream | accept | This is the documented residual risk per CONTEXT.md D-04. The Chrome "Sharing your screen" indicator is the user-facing signal; the operator chose to share. No code-level mitigation is possible (the API is supposed to capture the screen). Documented as accepted risk in 01-RESEARCH.md §"Security Domain". |
+| T-1-03 | Information Disclosure — captured stream contains other apps' content | `getDisplayMedia` stream | accept | This is the documented residual risk per CONTEXT.md D-04 — the operator opted into the "Sharing your screen" indicator as the cost of the broader capture coverage. The Chrome "Sharing your screen" indicator is the user-facing signal; the operator chose to share. No code-level mitigation is possible (the API is supposed to capture the screen). Documented as accepted risk in 01-RESEARCH.md §"Security Domain". |
 | T-1-NEW-03-01 | DoS — unbounded buffer growth on a stuck timestamp | `addChunk` + `trimAged` | mitigate | The trim function uses arrival timestamp; if a clock anomaly produces a negative or stuck `now`, the buffer is bounded above by SPEC §10 #9 (50 MB RAM ceiling) anyway. Defensive belt: the recorder also exposes `getBuffer().length` to SW so a healthchecker can observe growth. No active rate-limit needed for Phase 1. |
 </threat_model>
 
@@ -519,20 +524,20 @@ export class OffscreenLogger {
     this.context = context;
   }
 
-  private logWithLevel(level: 'log' | 'warn' | 'error', ...args: any[]) {
+  private logWithLevel(level: 'log' | 'warn' | 'error', ...args: unknown[]) {
     const timestamp = new Date().toISOString();
     console[level](`[OS:${this.context}] ${timestamp}`, ...args);
   }
 
-  log(...args: any[]) {
+  log(...args: unknown[]) {
     this.logWithLevel('log', ...args);
   }
 
-  warn(...args: any[]) {
+  warn(...args: unknown[]) {
     this.logWithLevel('warn', ...args);
   }
 
-  error(...args: any[]) {
+  error(...args: unknown[]) {
     this.logWithLevel('error', ...args);
   }
 }

.planning/phases/01-stabilize-video-pipeline/01-04-PLAN.md

@@ -2,7 +2,7 @@
 phase: 01-stabilize-video-pipeline
 plan: 04
 type: tdd
-wave: 1
+wave: 3
 depends_on: ["02", "03"]
 files_modified:
   - src/offscreen/recorder.ts
@@ -14,9 +14,9 @@ requirements_addressed:
 
 must_haves:
   truths:
-    - "On module import, the offscreen opens exactly one port via `chrome.runtime.connect({ name: 'video-keepalive' })` AND emits exactly one `OFFSCREEN_READY` message via `chrome.runtime.sendMessage`"
+    - "On module import, the offscreen opens exactly one port via `chrome.runtime.connect({ name: 'video-keepalive' })` AND emits exactly one `OFFSCREEN_READY` message via `chrome.runtime.sendMessage` — this is the long-lived port keepalive that replaces `chrome.alarms` for SW idle-timer resets (D-17)"
     - "Both the port-connect and the OFFSCREEN_READY send happen AFTER `chrome.runtime.onMessage.addListener` registration (Pattern 4 ordering)"
-    - "When the open port fires its registered `onDisconnect` listener, the module synchronously reconnects (Pitfall 4 mitigation)"
+    - "When the open port fires its registered `onDisconnect` listener, the module synchronously reconnects (Pitfall 4 mitigation; preserves the D-17 contract across the ~5 min port-lifetime cap)"
     - "The port emits a `{ type: 'PING' }` postMessage on a ≤ 25 s interval (RESEARCH.md Pattern 5)"
     - "Pre-emptive reconnect runs every 290 s (belt-and-braces against the ~5 min port-lifetime cap)"
     - "The port handles incoming `REQUEST_BUFFER` messages by responding with `{ type: 'BUFFER', chunks: <getBuffer()> }`"
@@ -51,10 +51,12 @@ a contract that survives SW unloads, port 5-min cap, and offscreen
 bootstrap races.
 
 Output: Updated `src/offscreen/recorder.ts` — only this single file is
-touched. Plan 04 is an isolated unit-of-work; Plan 05 picks up the
+touched. Plan 04 is an isolated unit-of-work; Plan 05 (the SW-side
-SW-side `onConnect` handler in parallel (Plan 04 runs alongside Plan 05 in
+`onConnect` counterparty) follows in wave 4 — Plan 04 lands first in wave 3
-Wave 1 with NO file overlap — Plan 04 owns offscreen-side, Plan 05 owns
+so the SW-side host has a well-defined offscreen-side contract to bind
-SW-side).
+against. Plan 04 owns the offscreen-side; Plan 05 owns the SW-side; the two
+files do not overlap (Plan 04 touches only `src/offscreen/recorder.ts`,
+Plan 05 touches only `src/background/index.ts`).
 </objective>
 
 <execution_context>

.planning/phases/01-stabilize-video-pipeline/01-05-PLAN.md

@@ -2,7 +2,7 @@
 phase: 01-stabilize-video-pipeline
 plan: 05
 type: execute
-wave: 2
+wave: 4
 depends_on: ["03", "04"]
 files_modified:
   - src/background/index.ts
@@ -14,13 +14,14 @@ requirements_addressed:
 
 must_haves:
   truths:
-    - "`src/background/index.ts` no longer contains `addVideoChunkFromBlob`, `cleanupVideoBuffer`, `setupKeepalive`, `loadChunkFromIndexedDB`, `openIndexedDB`, or any `chrome.alarms` reference (buffer ownership moves to offscreen per D-16/D-19)"
+    - "`src/background/index.ts` no longer contains `addVideoChunkFromBlob`, `cleanupVideoBuffer`, `setupKeepalive`, `loadChunkFromIndexedDB`, `openIndexedDB`, or any `chrome.alarms` reference — buffer ownership moves to offscreen per D-16/D-19, and the alarms-driven keepalive is DELETED per D-18 (it never actually reset the SW idle timer; the long-lived port does)"
-    - "`src/background/index.ts` no longer calls `chrome.tabCapture.getMediaStreamId` (D-01 amendment)"
+    - "`src/background/index.ts` no longer calls `chrome.tabCapture.getMediaStreamId` (D-01 amendment); video acquisition is now `getDisplayMedia` invoked from the offscreen module"
     - "`src/background/index.ts` no longer handles `VIDEO_CHUNK` or `VIDEO_CHUNK_SAVED` (deleted Message types in Plan 03)"
-    - "SW has an `onConnect` listener that filters `port.name === 'video-keepalive'` and validates `port.sender?.id === chrome.runtime.id` (T-1-04 mitigation)"
+    - "`src/background/index.ts` no longer contains any `chrome.tabs.onActivated` handler tied to the recording lifecycle (D-14: tab-switch re-attach is non-applicable under `getDisplayMedia`; D-15: operator tab-switching no longer interrupts recording, the buffer keeps filling regardless of active tab)"
+    - "SW has an `onConnect` listener that filters `port.name === 'video-keepalive'` and validates `port.sender?.id === chrome.runtime.id` (T-1-04 mitigation; this is the SW-side counterparty of the long-lived port keepalive per D-17)"
     - "SW has an `onMessage` `OFFSCREEN_READY` case that resolves a pending readiness Promise (Pattern 4 SW side)"
     - "SW's `SAVE_ARCHIVE` and `GET_VIDEO_BUFFER` handlers fetch the buffer via the port (`REQUEST_BUFFER` → wait for `BUFFER`) instead of holding their own `videoBuffer` array"
-    - "SW's `ensureOffscreen` uses `chrome.offscreen.Reason.DISPLAY_MEDIA` (not `USER_MEDIA`)"
+    - "SW's `ensureOffscreen` uses `chrome.offscreen.Reason.DISPLAY_MEDIA` (not `USER_MEDIA`) per D-02"
     - "SW's `onInstalled` listener calls `indexedDB.deleteDatabase('VideoRecorderDB')` once as a cleanup pass (RESEARCH.md Runtime State Inventory)"
     - "`npx tsc --noEmit` exits 0"
   artifacts:
@@ -49,9 +50,11 @@ ring-buffer state and helpers (now owned by offscreen per D-16), DELETE the
 chrome.alarms keepalive (D-18), DELETE the IndexedDB code path (D-19),
 DELETE the `chrome.tabCapture.getMediaStreamId` call (D-01 amendment),
 DELETE the `VIDEO_CHUNK` / `VIDEO_CHUNK_SAVED` message handlers (their
-message types were removed in Plan 03), and WIRE the SW-side `onConnect`
+message types were removed in Plan 03), DELETE any `chrome.tabs.onActivated`
-handler against the `'video-keepalive'` port that Plan 04 opens from the
+re-attach plumbing (D-14: not applicable under the new capture API; D-15:
-offscreen.
+operator tab switches no longer interrupt the recording), and WIRE the
+SW-side `onConnect` handler against the `'video-keepalive'` port that
+Plan 04 opens from the offscreen.
 
 Purpose: REQ-video-ring-buffer's data flow on export is `popup →
 SAVE_ARCHIVE → SW → REQUEST_BUFFER (via port) → offscreen → BUFFER (via
@@ -200,10 +203,12 @@ If `chrome.offscreen.Reason.DISPLAY_MEDIA` is NOT in the current `@types/chrome`
 
 (13) Delete `openIndexedDB` function — currently lines 507-520.
 
+(14) Verify no `chrome.tabs.onActivated` listener exists in the file (D-14 / D-15: tab-switch handling is non-applicable under the new capture API). Run `grep -n "chrome.tabs.onActivated" src/background/index.ts`. If the grep returns any hits, DELETE those lines (the entire listener callback block). If the grep returns nothing, log "D-14/D-15 satisfied: no tab-switch handler found in SW" in the task summary.
+
 After ALL these deletions, run `npx tsc --noEmit`. It MUST exit 0. If `VideoChunk` is reported as unused after the deletes, that indicates a function that needs it was inadvertently lost; STOP and audit.
   </action>
   <verify>
-    <automated>npx tsc --noEmit && [ $(grep -cE "addVideoChunkFromBlob|cleanupVideoBuffer|setupKeepalive|loadChunkFromIndexedDB|openIndexedDB|getMediaStreamId|chrome\.alarms" src/background/index.ts) -eq 0 ]</automated>
+    <automated>npx tsc --noEmit && [ $(grep -cE "addVideoChunkFromBlob|cleanupVideoBuffer|setupKeepalive|loadChunkFromIndexedDB|openIndexedDB|getMediaStreamId|chrome\.alarms|chrome\.tabs\.onActivated" src/background/index.ts) -eq 0 ]</automated>
   </verify>
   <acceptance_criteria>
     - `npx tsc --noEmit` exits 0
@@ -215,6 +220,7 @@ After ALL these deletions, run `npx tsc --noEmit`. It MUST exit 0. If `VideoChun
     - `grep -v '^#' src/background/index.ts | grep -c "getMediaStreamId"` returns 0
     - `grep -v '^#' src/background/index.ts | grep -c "VIDEO_CHUNK_SAVED"` returns 0
     - `grep -v '^#' src/background/index.ts | grep -c "chrome.alarms"` returns 0
+    - `grep -v '^#' src/background/index.ts | grep -c "chrome.tabs.onActivated"` returns 0 (D-14/D-15 mitigation)
     - `grep -c "DISPLAY_MEDIA" src/background/index.ts` returns 1
     - `grep -c "as any" src/background/index.ts` returns 0 (CLAUDE.md rule)
     - File line count reduced from 536 to roughly 380-400 lines (allow ±40)
@@ -435,7 +441,7 @@ Commit cadence: TWO commits.
 </verification>
 
 <success_criteria>
-- `src/background/index.ts` carries no buffer state, no alarms, no IndexedDB plumbing, no `tabCapture` calls
+- `src/background/index.ts` carries no buffer state, no alarms, no IndexedDB plumbing, no `tabCapture` calls, no `chrome.tabs.onActivated` re-attach plumbing
 - SW has `onConnect` handler matching the offscreen's port (Plan 04 counterparty)
 - SW has `OFFSCREEN_READY` handshake handler resolving a readiness Promise
 - T-1-04 mitigations in place on BOTH onConnect (sender + port name) and onMessage (sender)

.planning/phases/01-stabilize-video-pipeline/01-06-PLAN.md

@@ -2,12 +2,13 @@
 phase: 01-stabilize-video-pipeline
 plan: 06
 type: execute
-wave: 2
+wave: 5
-depends_on: ["03"]
+depends_on: ["03", "05"]
 files_modified:
   - vite.config.ts
   - offscreen/index.ts
   - offscreen/index.html
+  - src/background/index.ts
 autonomous: true
 requirements:
   - REQ-video-ring-buffer
@@ -16,11 +17,11 @@ requirements_addressed:
 
 must_haves:
   truths:
-    - "`vite.config.ts` no longer contains the `copy-offscreen` inline plugin (the 200+-line block including IndexedDB plumbing, codec fallback chain, and `mediaRecorder` shadow is GONE)"
+    - "`vite.config.ts` no longer contains the `copy-offscreen` inline plugin — the entire 200+-line block including IndexedDB plumbing, codec fallback chain, and `mediaRecorder` shadow is GONE per D-08 (orphan dead-code deletion plus inline-plugin deletion in lockstep)"
-    - "`vite.config.ts` declares the offscreen entry via `rollupOptions.input` per RESEARCH.md Example B"
+    - "`vite.config.ts` declares the offscreen entry via `rollupOptions.input` per RESEARCH.md Example B; crxjs picks up the new TS entry through the HTML reference and the runtime path remains `offscreen/index.html` resolvable via `chrome.runtime.getURL(...)` per D-07"
-    - "Top-level `offscreen/index.ts` is DELETED (dead code per audit P2 #18)"
+    - "Top-level `offscreen/index.ts` is DELETED (dead code per audit P2 #18; D-08 explicit DELETE target)"
-    - "Top-level `offscreen/index.html` is DELETED (replaced by the crxjs-managed `src/offscreen/index.html` from Plan 03)"
+    - "Top-level `offscreen/index.html` is DELETED (REPLACED by the crxjs-managed `src/offscreen/index.html` from Plan 03, per D-07 — the runtime URL semantics survive the source-tree move)"
-    - "`npm run build` exits 0; `dist/` contains a bundled offscreen HTML at the path the SW's `chrome.runtime.getURL` argument expects"
+    - "`npm run build` exits 0; `dist/` contains a bundled offscreen HTML at the path the SW's `chrome.runtime.getURL` argument expects (D-07 runtime-path contract)"
     - "No mention of `VideoRecorderDB`, `openIndexedDB`, `chromeMediaSource`, or `copy-offscreen` remains in `vite.config.ts`"
   artifacts:
     - path: "vite.config.ts"
@@ -106,8 +107,8 @@ const url = chrome.runtime.getURL('offscreen/index.html');
 
 | Threat ID | Category | Component | Disposition | Mitigation Plan |
 |-----------|----------|-----------|-------------|-----------------|
-| T-1-NEW-06-01 | Tampering — string-injected code via inline plugin | `vite.config.ts:13-216` `copy-offscreen` plugin's `this.emitFile({ source: \`<JS-as-string>\` })` | mitigate | DELETE the entire inline plugin. The replacement is a `src/offscreen/recorder.ts` real module + a `src/offscreen/index.html` declared in `rollupOptions.input`. No more long template-literal JS in `vite.config.ts`. Grep gate: `grep -v '^#' vite.config.ts \| grep -c "this.emitFile"` returns 0. |
+| T-1-NEW-06-01 | Tampering — string-injected code via inline plugin | `vite.config.ts:13-216` `copy-offscreen` plugin's `this.emitFile({ source: \`<JS-as-string>\` })` | mitigate | DELETE the entire inline plugin (D-08). The replacement is a `src/offscreen/recorder.ts` real module + a `src/offscreen/index.html` declared in `rollupOptions.input`. No more long template-literal JS in `vite.config.ts`. Grep gate: `grep -v '^#' vite.config.ts \| grep -c "this.emitFile"` returns 0. |
-| T-1-NEW-06-02 | Tampering — orphaned root-level offscreen | `offscreen/index.ts` + `offscreen/index.html` (dead code) | mitigate | DELETE both files. After this plan, a `find offscreen/ -type f` produces no output. Grep gate: `[ ! -d offscreen ]` exits 0. |
+| T-1-NEW-06-02 | Tampering — orphaned root-level offscreen | `offscreen/index.ts` + `offscreen/index.html` (dead code) | mitigate | DELETE both files (D-07 / D-08 — runtime path is preserved via the crxjs-managed `src/offscreen/index.html`). After this plan, a `find offscreen/ -type f` produces no output. Grep gate: `[ ! -d offscreen ]` exits 0. |
 </threat_model>
 
 <tasks>

.planning/phases/01-stabilize-video-pipeline/01-07-PLAN.md

@@ -2,7 +2,7 @@
 phase: 01-stabilize-video-pipeline
 plan: 07
 type: execute
-wave: 3
+wave: 6
 depends_on: ["05", "06"]
 files_modified:
   - tests/fixtures/last_30sec.webm
@@ -139,7 +139,7 @@ The full Phase-1 Mokosh build:
 - Archive is delivered via `chrome.downloads`
   </what-built>
   <verify>
-    <automated>which ffprobe && ffprobe -v error -f matroska -i tests/fixtures/last_30sec.webm; test $? -eq 0</automated>
+    <automated>test -f tests/fixtures/last_30sec.webm && which ffprobe && ffprobe -v error -f matroska -i tests/fixtures/last_30sec.webm; test $? -eq 0</automated>
   </verify>
   <how-to-verify>
 **Pre-flight (automated, do these in a bash shell before opening Chrome):**

.planning/phases/01-stabilize-video-pipeline/01-RESEARCH.md

@@ -1017,10 +1017,11 @@ stream.getTracks().forEach((track) => {
 **If this table contains items:** The planner should treat them as
 candidates for user verification during `/gsd-plan-phase` review.
 
-## Open Questions
+## Open Questions (RESOLVED)
 
 1. **Will `MediaRecorder.start(2000)` produce ffprobe-clean WebM on a
    typical screen-cap?**
+   - **RESOLVED:** Cluster-boundary alignment is resolved by the D-12 ffprobe acceptance gate (enforced in Plan 03 Task 2 verify path + Plan 07 Task 1) and the D-13 restart-segments fallback (pre-staged as a commented skeleton in `src/offscreen/recorder.ts` per Plan 03; activated by re-plan after Plan 07 if the gate fails).
    - What we know: Cluster boundaries align with keyframes; Chrome
      keyframes appear every ~3-5 s by default (vp9 `kf_max_dist=100` on
      a 30 fps stream); timeslice does NOT force keyframes.
@@ -1034,6 +1035,7 @@ candidates for user verification during `/gsd-plan-phase` review.
      this in the success criteria.
 
 2. **Does the 5-minute port lifetime kill the recording session?**
+   - **RESOLVED:** Plan 04's 290 s pre-emptive reconnect logic plus the synchronous onDisconnect → connectPort reconnect path mitigate the cap whether it applies to port lifetime or SW lifetime; either way the offscreen reconnects within seconds and the buffer is unaffected.
    - What we know: Multiple corroborating community sources cite a ~5
      minute hard cap on long-lived ports.
    - What's unclear: Whether the cap applies to *port lifetime* (the
@@ -1045,6 +1047,7 @@ candidates for user verification during `/gsd-plan-phase` review.
      reconnect is still harmless.
 
 3. **What's the exact crxjs path-emit behavior for the offscreen entry?**
+   - **RESOLVED:** Plan 06 Task 2 performs runtime verification — runs `npm run build`, inspects `dist/` for whichever of `dist/src/offscreen/index.html` or `dist/offscreen/index.html` was emitted, then edits `src/background/index.ts`'s `chrome.runtime.getURL(...)` argument to match (this is why Plan 06 now lists `src/background/index.ts` in files_modified per the iteration-1 dependency-correctness fix).
    - What we know: The discussion #919 working answer uses
      `input: { offscreen: 'src/offscreen/offscreen.html' }` and SW
      fetches `chrome.runtime.getURL('src/offscreen/offscreen.html')`.