Allsup Leads Ssdi

Allsup Leads — SSDI

The monthly SSDI/SSI claimant batch for the Allsup work (via Pat Sullivan). This skill IS the proven process — it replaces the /refine skill for this vertical. /refine (Sherlock confirm + blue-healer resolve) is the WRONG path here: it collapses the batch and yields ~0 contacts. Use this instead.

The product: handle + distress quote + ≥1 reachable public profile. Graded on REACHABILITY, not contact data. Allsup owns all outreach — we never contact anyone.

Quick start

scripts/run.sh            # pull + tier + rename + build the book (default 30 days)
scripts/run.sh --days 45  # wider window for more volume

Then publish the book and draft the email (steps 3–4 below).

Workflow

1. Pull + tier + build book — scripts/run.sh. This mines ssdi-work-fear on Reddit + X (fresh, no ledger), tiers by urgency, writes ~/Desktop/ALLSUP-LEADS-<date>.{csv,md}, and builds .book/index.html. Report the tier counts (best/better/good) it prints.
2. Verify the CSV has real quotes + profiles (don't ship empty tiers): python3 -c "import csv; r=list(csv.DictReader(open('$HOME/Desktop/ALLSUP-LEADS-<date>.csv'))); print(len(r),'rows; with quote:',sum(1 for x in r if x['quote'].strip()))"
3. Publish the book (permanent, forwardable — Pat sends it to Allsup): ~/.codex/skills/here-now/scripts/publish.sh <skill>/.book --client codex Uses Brent's saved here.now key (permanent). Grab the site_url.
4. Draft the email to Pat (pat@contatta.com) via the Gmail tool — DRAFT only, never send; Brent reviews and sends. Use the template in EMAIL.md with this run's counts + book URL. Tell Brent to drag ALLSUP-LEADS-<date>.csv in as the attachment before sending.

The rules that make this work (hard-won — do not violate)

• Reddit is the densest source and is KEYLESS. Never exclude it. Oxylabs only speeds it up.
• Grade on reachability, not contact data. A Reddit/X profile IS a reachable channel.
• NEVER run the heavy resolver (blue-healer maigret / email-finder / SMTP). On pseudonymous handles it yields ~0, burns API calls, and creates PII/compliance exposure. The lean pull IS the product. (This is why /refine is wrong for SSDI.)
• No emails/phones. They aren't findable for these handles and aren't the deliverable. The profile is the channel — Allsup DMs it. Say this plainly to Pat (see EMAIL.md).
• SSDI only. Other Allsup lead types are a different, non-sensitive vertical — build a separate skill for them; don't add them here.

Tiers (encode claim urgency)

• best — denied / appeal / reconsideration / hearing / overpayment / can't work
• better — applying / pending / waiting / nervous
• good — work-incentive / Ticket-to-Work / "will working cost my benefits"

Paths & deps

• Proven pull: ~/Desktop/Datamine/blue-healer/Datamine-repo/scripts/allsup_pull.py
• Datamine bin: ~/Datamine/.venv/bin/datamine (auto-set as DATAMINE_BIN)
• Publisher: ~/.codex/skills/here-now/scripts/publish.sh (Brent's key already saved)
• Canonical background: ALLSUP_RUNBOOK.md in the Datamine-repo; memory project_ssdi_allsup_proven_pipeline_runbook.

Codex Runtime

Resolve sibling Codex skills from installed skill roots (including ~/.codex/skills) or the current collection; do not assume a legacy skill root.

Never expose or print secret, credential, or token values.

Mandatory dependencies:

• Datamine environment
• last30days data-source credentials
• here.now publishing credentials

Preflight each dependency using MCP/app capability discovery, CLI availability/version checks, read-only filesystem or Git checks for repositories, and provider auth-status commands without printing secrets, credentials, or tokens. If any mandatory dependency is unavailable, stop and report a concise blocked state naming the missing dependency and the next action needed.