# Wake Report Copy QA Card

Status: `wake_report_copy_qa_card_manual_report_only_closed_until_human_yes`

This is a manual-report-only QA card for the final Sonic-Forage wake report copy. It sits after the final paste card and checks that every line still matches the current repo, verifier output, source map, Pages/proof evidence, and stream/provider boundary before an awake operator uses it.

## Human approval question

Do you approve using the Sonic-Forage wake report copy QA card as a manual final wording check after rerunning the verifier stack and confirming the exact report lines remain true?

## Exact commands to rerun first

```bash
PYTHONDONTWRITEBYTECODE=1 python3 scripts/verify_site.py
python3 -m json.tool docs/data/wake-report-copy-qa-card.json >/dev/null
git diff --check
git status --short --branch
```

Expected highlights:

- `VERIFY OK afterparty explainer site`
- JSON parses cleanly
- `git diff --check` exits with no output
- Git status matches the exact local/pushed state being reported

## Five copy QA checks

1. **Current repo state:** the final report lines cite only the current checkout, current verifier output, and current Pages/proof paths. If push or Pages status is not verified, say local-only or pending instead.
2. **Closed gate scan:** every line preserves no posting, no upload, no stream, no outreach, no payment, no provider credential check, no GPU/training, no secret exposure, no voice-to-shell, and no cron mutation.
3. **Source-map alignment:** every proof claim maps to `docs/reports/WAKE_REPORT_SOURCE_MAP.md`, `docs/reports/LAUNCH_PROOF_INDEX.md`, `docs/reports/PAGES_VERIFICATION_RECEIPT.md`, or the one-screen readout.
4. **Stream/provider boundary:** any stream or media wording distinguishes local media/proof health from live RTMP, Kick, Twitch, OBS, chat, or credential approval.
5. **Human next-step boundary:** the report ends with zero or one human-approved next route from `docs/launch/PROOF_HUB_NEXT_DECISION_COMPASS.md`, or keeps building local proof.

## 45-second copy QA route

1. Run `PYTHONDONTWRITEBYTECODE=1 python3 scripts/verify_site.py`; look for `VERIFY OK afterparty explainer site`.
2. Run `git diff --check`; expect no output.
3. Open `docs/reports/WAKE_REPORT_FINAL_PASTE_CARD.md` and remove any stale line before delivery.
4. Open `docs/reports/WAKE_REPORT_SOURCE_MAP.md` for every proof, Pages, dataset, media, or stream-status sentence.
5. Open `docs/streaming/STREAM_LOOP_LOCAL_HEALTH_RECEIPT.md` before saying anything about stream-loop/media health; keep provider/session/credential claims out unless separately verified by a human-approved check.
6. Stop before posting, uploading, streaming, billing, contacting, checking provider credentials, running voice-to-shell, or mutating cron.

## Safe wording

- `manual-report-only copy QA`
- `verified against repo-local proof paths`
- `closed until human approval`
- `local media proof, not live provider approval`
- `no external action performed`

## Do-not-say lines

- Do not say posted the launch.
- Do not say uploaded the video.
- Do not say started the stream.
- Do not say buyers were contacted unattended.
- Do not say revenue was made.
- Do not say OpenAI affiliation exists.
- Do not say the dataset is public.
- Do not say GPU training started.
- Do not say provider credentials were checked.
- Do not say cron was updated.

## Closed gates

- public_posting=false
- social_scheduling=false
- youtube_upload=false
- caption_upload=false
- publishes_stream=false
- records_audio=false
- uploads_private_media=false
- outreach=false
- payments_or_invoices=false
- claim_revenue=false
- claim_openai_affiliation=false
- public_dataset_release=false
- hf_upload=false
- starts_gpu=false
- starts_paid_api=false
- downloads_models=false
- starts_training=false
- provider_credential_check=false
- voice_to_shell=false
- secret_exposure=false
- mutates_cron=false

## Blocked without approval

- Posting or scheduling social copy.
- Uploading YouTube videos, captions, Shorts, thumbnails, datasets, or private media.
- Starting Twitter Spaces, livestreams, OBS, RTMP, Twitch, Kick, recordings, or chat integrations.
- Sending outreach, DMs, email, forms, buyer follow-ups, or public announcements.
- Creating payment links, checkout, invoices, wallets, paid promotion, or revenue claims.
- Claiming customers, public launch, provider approval, platform affiliation, or OpenAI affiliation.
- Checking provider credentials or printing secret-bearing environment details.
- Starting GPU, paid API, model download, Modal job, training, or evaluation work.
- Running voice-to-shell or live-provider automations.
- Creating, updating, pausing, removing, or scheduling cron jobs.

## Proof paths

- `docs/data/wake-report-copy-qa-card.json`
- `docs/reports/WAKE_REPORT_FINAL_PASTE_CARD.md`
- `docs/reports/WAKE_REPORT_SOURCE_MAP.md`
- `docs/reports/PAGES_VERIFICATION_RECEIPT.md`
- `docs/streaming/STREAM_LOOP_LOCAL_HEALTH_RECEIPT.md`
- `docs/launch/PROOF_HUB_NEXT_DECISION_COMPASS.md`
- `scripts/verify_site.py`
- `docs/index.html`