Fix re-ingest churn: exclude the Claude status file from the re-ingest signature (stops needless re-summarization)

You are being dispatched from flower Brief #134: Fix re-ingest churn: exclude the Claude status file from the re-ingest signature (stops needless re-summarization)

Recall pointer:
- Use recall_brief with id 134 for the full folder if you need provenance.

Target:
- project: flower (/Users/mikeferrara/Documents/code/flower)
- branch: choose an appropriate branch
- worktree: not specified
- kind: fresh

Current brief spec:
(no spec yet)

This is a direct request, not a fully-specced plan. If it's clear, resolve it. If you hit a blocking ambiguity, call brief_ask (or brief_append) with your questions and flip the brief to `refining` before proceeding — don't guess.

Recent/key trace events:
[1] participant_joined flower-ops: (no body)
[2] note_added flower-ops: **Operator (Mike) reported 2026-07-03** (flower sessions "going backwards"; wasteful re-summarization). Root-caused by flower-ops. Canonical fix-spec: Solo scratchpad **#1053**; tracking todo **#693**. Elevating to a brief so it gets dispatched (this is the fix that STOPS the re-summarizations).

## Problem (efficiency/correctness)
`ClaudeAdapter::fileSignature()` builds the re-ingest change-detection signature over **[transcript file + the Claude session status file `~/.claude/sessions/<n>.json`]** (`path:mtime:size` each), and **subagents inherit the PARENT's status file**. That status file is rewritten every few seconds for any LIVE session, so the signature changes with **zero new transcript messages** → `ScansHarnessSessions` re-dispatches `IngestSession` → re-segment → re-summarize → sessions flap `indexed → summarized → indexed` (wasted deepseek calls). Hits every long-running daemon session + all its historical subagent transcripts on every `flower:watch`.

**Confirmed corpus-wide** (not flower-only): a single scan produced a 57-session parsed+summarized batch — 56/57 = churn (old content re-ingested <1h), 32/57 subagents; across tdr-thedarkroom(61)=34, sierra-firewood(49)=20, flower, swinglist, enmbook. Smoking gun: session #706 sig = static transcript (mtime 06-29) | status file 96293.json mtime=today.

## Fix (operator-endorsed: "ensure new messages before re-ingest")
- **(A)** Change-detect on the **transcript only** (mtime+size / content_hash), not the volatile status file; subagents must NOT inherit the parent's status file; a status-only change refreshes cheap metadata WITHOUT re-segment/re-summarize.
- **(B)** Belt-and-suspenders in `IngestSession`: short-circuit when the transcript `content_hash` is unchanged (no new messages) → refresh + return, no re-summarize. Ensure that content_hash excludes the status file.
- Check `CodexAdapter::fileSignature` for the same volatile-status pattern.

## Files
app/Harness/Adapters/ClaudeAdapter.php, AbstractHarnessAdapter.php, app/Console/Commands/Concerns/ScansHarnessSessions.php, app/Jobs/IngestSession.php (+ CodexAdapter.php).

## Verify
`flower:watch` reports `skipped_unchanged` for static subagent transcripts (dispatched≈0 when no transcript changed); no session flaps indexed→summarized without new events; genuine transcript appends still re-ingest+re-summarize; session-ended detection not regressed. `php artisan test` green; pint clean.

Full detail + repro queries in scratchpad #1053.
[3] participant_joined operator:mike: (no body)
[4] status_change operator:mike: (no body)
[5] link_added flower-ops: (no body)

Recommended linked context:
{
    "todos": [],
    "scratchpads": [
        {
            "id": 333,
            "solo_scratchpad_id": "1005",
            "name": "flower-ops — triage log",
            "archived": false,
            "revision": 208
        }
    ]
}

Execution notes:
- Treat the brief as the source of truth.
- Keep work scoped to this dispatch request.
- Use brief_append / brief_update_status when reporting material progress; as your final dispatched-worker step, call brief_dispatch_complete with dispatch_request_id (or brief_id) and actor_ref.
- Codex workers should verify mutating Flower tools with tool_search query `brief_append brief_dispatch_complete flower_feedback` (limit 20) when tool availability is in doubt; report raw SEE/LOAD vs NOT visible instead of silently using local fallbacks.
- Add a git commit trailer `Brief: #134` to every commit for this brief so flower can exact-link commits back to the brief.