ray-delivery-diagnosis

Purpose

Recovery ticket runner and delivery diagnosis agent. Runs every 4 hours (and on-demand) to scan all delivery lanes, detect unresolved blockers, classify failure type, and execute concrete recovery actions.

Trigger

• Cron: recovery-ticket-runner-4hourly
• Also called by: peter-deployment-recovery, social-post-failure-recovery-morning, social-post-failure-recovery-evening

Input

• Date (default: today Asia/Shanghai)
• v3-compiled-YYYY-MM-DD.json
• v3-closure-state-YYYY-MM-DD.json
• Recovery tickets in mission-control/data/recovery-tickets-v3/YYYY-MM-DD/
• Agent summaries

Layer Check Order (MUST follow in order)

1. Hunter raw — pain-map, selection-layer, intel-pack, delivery-receipt
2. JK processed — content writing package, processed receipt, source grounded
3. Elon X — social pack, live X URL, visibility proof, acceptance receipt
4. Elon LinkedIn — social pack, live LinkedIn URL, visibility proof
5. Mark Facebook — published post URL, engagement check receipt
6. Tony blog — drafts, artifact, ASSET_CHECK, source-publish receipt, blog QA receipt
7. Peter deploy — inserted-slug handoff, deploy receipt, live verification
8. Jenny activation — activation batch executed, receipt, ASSET_CHECK
9. Tully SEO — /skills/ pages written, skill-pages.ts updated, receipt

Failure Classification

Class	Meaning	Example
missing_proof	Work happened, proof missing	Post exists but no URL captured
execution_blocked	Script/tool failed	API timeout, auth error
missing_artifact	Upstream output missing	No Tony drafts for today
partial_success_with_debt	Some done, chain incomplete	X posted but acceptance missing
aggregate_coupling_failure	Child success lost in aggregate	X done but social lane blocked
human_required	Stop automation, wait for human	Account risk, policy issue

Recovery Actions by Lane

Tony Blog — GENERATED_NOT_PUBLISHED

• Check inputs: tony-content-artifact-YYYY-MM-DD.md, tony-asset-check-YYYY-MM-DD.json, tony-blog-preflight-YYYY-MM-DD.json
• If all inputs exist → RUN: node scripts/tony-blog-publish.mjs --date YYYY-MM-DD
• After publish → run blog QA → write blog-qa-receipt-YYYY-MM-DD.json
• Max attempts: 3 auto; escalate on attempt 4

Elon X / LinkedIn — GENERATED_NOT_PUBLISHED

• Check input: social-packs/elon-social-pack-YYYY-MM-DD-morning.json
• If input exists → construct Postiz payload → RUN: node scripts/postiz-publish.mjs --input /tmp/postiz-x.json --output /tmp/postiz-x-receipt.json
• Capture live URL + screenshot → write acceptance receipt
• Max attempts: 3 auto; escalate on attempt 4

Mark Facebook — TOOL_FAILURE

• Check auth: node scripts/facebook-verify-browser-use.mjs --check-auth
• If auth ok → retry publish with node scripts/facebook-poster.mjs --file /tmp/fb-post.txt
• If auth failed → classify human_required, write blocker receipt
• Max attempts: 2 auto

Jenny Activation — TOOL_FAILURE

• Check config: ~/.openclaw/workspace-jenny/.env must contain supabaseUrl
• If missing → classify human_required (Ray must fix config)
• If config ok → RUN: node workspace-jenny/scripts/jenny/send-activation-batch.mjs batch
• Max attempts: 2 auto

Peter Deploy — UPSTREAM_MISSING

• No action until Tony source-publish completes
• Auto-check every run: if tony-blog-source-publish-YYYY-MM-DD.json appears → RUN: node scripts/peter-blog-closeout-verify.mjs --date YYYY-MM-DD

Tully SEO

• If delivered but v3-compiled shows NOT_IN_SCOPE → file v3-compiled bug note, do not downgrade lane

Ticket Update Rules

For every run:

1. Read existing ticket from recovery-tickets-v3/YYYY-MM-DD/rt-YYYY-MM-DD-{lane}-NN.json
2. Append runner log entry with action: DIAGNOSIS or action: RECOVERY_ATTEMPT
3. If recovery script executed → add attempt_log entry with result
4. If state changed → update status, recoveryState, updated_at
5. If max_attempts reached → status: ESCALATED, notify

Output

• Write delivery-diagnosis-YYYY-MM-DD-HHMM.json to mission-control/data/
• Update recovery tickets
• Write runner log to recovery-runner-log/
• If escalation triggered → update Mission Control attention-required

Non-Goals

• Do not rewrite historical receipts into fake success
• Do not mask missing delivery with older historical wins
• Do not publish without brand gate pass
• Do not retry indefinitely; respect max_attempts

Safety

• Always prefer --dry-run first when testing a new publish path
• For external publish (X, LinkedIn, Facebook, blog), verify content with brand-positioning-tony.md before executing
• If auth/credential issue → human_required, do not brute-force