📌 Active Projects from MEMORY.md
Workspace start surface: dashboards/command-center.html is the generated open-first index and, as of Jun 12, now routes into the fresh Friday WFH package instead of the stale Jun 6 weekend board. Current active package: dashboards/friday-wfh-command-center-jun12.json, dashboards/friday-wfh-command-center-jun12.html, dashboards/command-center-index-jun12.json, dashboards/friday-wfh-launch-brief-jun12.md, and dashboards/friday-wfh-proof-log-jun12.md. That package deliberately grounds the first-click path in current Jun 11/12 work on disk: NYC PR / healthcare Step-5 verification first, previous-PELS draft lane second, Circle submit-proof queue third, and DOAC attribution repair fourth. The Jun 6 weekend→Monday package and earlier Friday/weekend boards are now reference context only. Integrity guardrail stays dashboards/check-command-center-links.js → dashboards/command-center-link-status.json; the Jun 12 routing refresh verified clean link resolution on disk.
PanArgent BD: highest-priority workstream. Canonical repo: panargent/bd-repo/. Current May 29/31 restart/research additions are panargent/bd-repo/notes/may29-fee-motion-owner-date-card.md, panargent/bd-repo/notes/may30-mighty-union-fee-motion-hook-card.md, panargent/bd-repo/notes/may31-monday-fee-motion-proof-ledger.md, panargent/bd-repo/research/deal-briefs/future-today-ib-deal-brief-may29.md, and the Monday June 1 fee-motion stack package (panargent/bd-repo/data/monday-fee-motion-jun1{,-pipeline}.json, panargent/bd-repo/scripts/build-monday-fee-motion-stack-jun1.js, panargent/bd-repo/monday-fee-motion-stack-jun1.html, panargent/bd-repo/action-packets/monday-fee-motion-packet-jun1.md, panargent/bd-repo/comms-kits/monday-fee-motion-comms-kit-jun1.html, panargent/bd-repo/launchpads/monday-fee-motion-launchpad-jun1.html, panargent/bd-repo/imports/2026-06-01-panargent-monday-fee-motion-stack-june-1-*, and panargent/bd-repo/qa/monday-fee-motion-stack-jun1-qa.md); the June 2 target-lane truth layer now matters too: panargent/bd-repo/scripts/build-target-lane-truth-dashboard.js, panargent/bd-repo/data/target-lane-truth-dashboard.json, panargent/bd-repo/dashboards/target-lane-truth-dashboard.html, and panargent/bd-repo/notes/jun2-target-lane-truth-dashboard-note.md scan saved target-lane notes for claim-vs-proof risk so packaged lanes do not get mistaken for compliant deliverables. The June 2 seller-motion gate now matters too: panargent/bd-repo/notes/jun2-binary-motion-gate-card.md defines the exact Outward → Mighty Union → Matter order plus the timestamped external-touch proof standard required before any motion claim counts as real. The June 3 fee-motion proof layer now matters too: panargent/bd-repo/scripts/build-fee-motion-proof-dashboard.js, panargent/bd-repo/data/fee-motion-proof-dashboard.json, panargent/bd-repo/dashboards/fee-motion-proof-dashboard.html, and panargent/bd-repo/notes/jun3-fee-motion-proof-dashboard-note.md separate concrete saved lane lines, gate-cleared states, and qualifying external-motion proof so a polished restart stack does not masquerade as seller progress. The same-day Mighty Union live-step gate now also lives at panargent/bd-repo/notes/jun3-mighty-union-first-touch-proof-card.md; the approval-ready outreach packet is panargent/bd-repo/notes/jun3-overnight-mighty-union-touch-pack.md, and panargent/bd-repo/research/deal-briefs/molekule-consulting-ib-deal-brief-jun3.md is the new June 3 research-stage healthcare / biopharma-intelligence brief. These are still prep/research artifacts only, not outreach proof. The May 31 Miami correction layer still matters: panargent/bd-repo/targets/miami-all-subsectors-may27/miami-all-subsectors-draft-post-exclude-106-summary-may31.md, panargent/bd-repo/targets/miami-all-subsectors-may27/miami-all-subsectors-draft-research-queue-99-summary-may31.md, and panargent/bd-repo/targets/miami-advertising-may31/research-summary.md preserve Hunter's count correction while keeping the lane honestly labeled draft/not compliant. The May 27 support layer (panargent/bd-repo/notes/may27-fee-motion-decision-sheet.md, panargent/bd-repo/targets/miami-all-subsectors-may27/miami-all-subsectors-target-list-summary-may27.md, and panargent/bd-repo/targets/restoration-service-lines-final-may27/restoration-service-lines-final-pels-may27.md) still stands behind them. All of these are restart/research support only, not outreach proof or compliant final-deliverable proof. PanArgent pushes are healthy; the old sync trap was assuming one repo-scoped PAT could cover rudy-memory too.
Job Search: active salary-upgrade lane. Latest lane-local proof artifacts are job-search/may29-official-submit-proof-ladder.md and job-search/jun3-circle-capital-one-official-proof-gate.md, which sharpen the exact Circle → Capital One submit-or-blocker order and the employer-surface proof standard; job-search/may27-submit-proof-card.md remains the earlier lane-local reset. Ground live role/material count from job-search/apr25-salary-upgrade-status.md and job-search/. Current execution artifacts: job-search/warm-intro-priority-list-may5.md, job-search/warm-intro-forwardable-packet-may6.md, job-search/top-4-packaged-application-order-may6.md, the compressed Friday cockpit job-search/friday-may8-salary-upgrade-cockpit.md, the send-speed layer job-search/friday-may8-warm-intro-send-sheet.md, the application tailoring grid job-search/top-4-role-tailoring-grid-may7.md, the last-mile submit pack job-search/friday-may8-submit-links-pack.md, the verification layer job-search/application-link-truth-report-may9.md, the quick reset sheet job-search/monday-may11-application-priority-reset.md, the clean Monday submit sheet job-search/monday-may11-direct-apply-sheet.md, the warm-path Monday sequence job-search/monday-may11-network-first-outreach-sheet.md, the truth-backed cockpit job-search/monday-submit-cockpit-may11.html, the direct Circle endpoint verdict job-search/circle-apply-surface-truth-may12.md, the current Friday-ranked operator sheet job-search/thursday-may14-submit-ladder.md, the proof gate job-search/friday-may15-submit-proof-sheet.md, the Monday closer job-search/monday-circle-submit-closer-may16.md, the proof-capture helper job-search/circle-submit-proof-capture-card-may17.md, the verify-first operator note job-search/circle-submit-truth-refresh-may18.md, the fresh lane-local restart dashboard job-search/job-search-command-center-may19.html, the paired short operator memo job-search/tuesday-salary-upgrade-launch-brief-may19.md, the stricter job-search/circle-capital-one-verify-first-stack-may19.md, the May 21 blocker compressor job-search/may21-submit-proof-gap-card.md, and the Tuesday-first post-holiday proof card job-search/may23-tuesday-submit-proof-card.md. Treat missing or questionable direct application URLs inside that stack as real blockers, not “good enough” placeholders, and do not use the stale March dashboards/job-command-center.html as a current start surface.
DOAC / diaryofceo.online: deploy from doac/diaryofceo-site/ to Cloudflare Pages project diaryofceo-site. Custom domain + homepage + sample pages + ads.txt were verified live again on May 4 with AdSense installed; the blocker is distribution/monetization throughput now, not domain downtime. On May 5 a local rollout script added affiliate reading blocks across 18 high-intent pages; QA/support artifacts now live in doac/doac-affiliate-rollout-browser-qa-may5.md and doac/doac-reddit-affiliate-launch-pack-may5.md. Decision-pack artifacts now include revenue-products/doac-pages-deploy-prestage-may6.md, doac/doac-affiliate-rollout-approval-packet-may6.md, doac/doac-affiliate-rollout-friday-qa-runbook-may7.md, source-only precheck doac/doac-affiliate-priority-pages-static-precheck-may7.md, the blank worksheet doac/doac-affiliate-browser-qa-scorecard-may8.md, the recovery distribution draft doac/doac-reddit-relaunch-pack-may8.md, the source-level preflight helper doac/doac-affiliate-browser-qa-harness-may9.{js,md}, the safer discussion-first relaunch variants in doac/doac-reddit-relaunch-3-post-pack-may9.md, the ranked QA queue doac/doac-affiliate-browser-qa-priority-queue-may10.md, the live-browser blocker artifact doac/doac-affiliate-browser-qa-pass-may10-night.md, the deploy-source diagnosis doac/doac-live-affiliate-mismatch-diagnosis-may11.md, the authoritative path memo doac/doac-authoritative-deploy-path-may12.md, the deploy-truth audit doac/doac-affiliate-deploy-truth-may12.{md,json}, the rerunnable audit script doac/scripts/audit-affiliate-deploy-truth.mjs, the rollout-only staging memo doac/doac-affiliate-rollout-commit-manifest-may12-night.md, the HOLD-gated distribution prep doac/doac-discussion-first-launch-copy-pack-may12-night.md, the May 13 reopen artifact doac/doac-affiliate-live-strike-checklist-may13-night.md, the tighter Friday operator pack doac/doac-affiliate-friday-deploy-pack-may14-night.md, the proof gate doac/doac-live-rollout-proof-card-may15.md, the smaller-scope deploy card doac/doac-affiliate-smallest-safe-commit-card-may16.md, the read-only staging helper doac/scripts/prepare-affiliate-rollout-stage.mjs, the read-only live-check helper doac/scripts/check-affiliate-rollout-live.mjs, the tighter execution path doac/doac-affiliate-live-proof-run-card-may17.md, the read-only local verifier doac/scripts/verify-affiliate-rollout-local.mjs, the commit-scope blocker note doac/doac-affiliate-commit-delta-proof-may18.md, the approved-file validator doac/scripts/print-affiliate-rollout-approved-files.mjs, the read-only scope-control artifact doac/doac-affiliate-approved-commit-manifest-may19-night.md, the May 20 whitespace gate doac/doac-approved-rollout-whitespace-gate-may20.md, the May 20 Friday-use commit proof card doac/doac-approved-rollout-commit-proof-card-may20-night.md, the new read-only preflight helper doac/scripts/run-approved-rollout-preflight-may23.mjs, its operator note doac/may23-approved-rollout-preflight-runbook.md, and the compact approved-scope gate map doac/may23-approved-rollout-proof-ladder.md. A separate root-level site audit now also lives at diaryofceo-site/PHASE-1-EDITORIAL-REDESIGN-AUDIT-2026-05-20.md; that is audit-only and not a deploy-path artifact. Governing truth: production currently follows committed main on doac/diaryofceo-site/ and the live top pages still match HEAD, not the dirty local worktree, so the affiliate rollout is blocked on commit/push/deploy truth rather than copywriting. New May 25 gate truth: the approved-file whitespace blocker is now cleared locally via doac/scripts/fix-approved-rollout-whitespace-may25-night.mjs, doac/may25-approved-rollout-tuesday-proof-sheet.md, and dashboards/tuesday-morning-operator-card-may26.md; staging/deploy work is still blocked on Hunter-owned approved-scope git motion plus post-deploy browser proof. On May 26 the DOAC newsletter lane also produced one real distribution event: doac/newsletters/edition-17-send-result-may26.json shows Edition 17 (Naval Ravikant) sent to 21 unique recipients with 21 sent / 0 failed, supported by edition-17-naval-ravikant.html, send-test-edition-17-may26.js, send-edition-17-to-subscribers-may26.js, edition-17-sent-log.json, and edition-17-distribution-pack-may26-night.md. May 27 follow-ons doac/newsletters/edition-17-revenue-bridge-may27.md and doac/newsletters/edition-17-social-launch-sequence-may27-night.md preserve the next monetization/distribution logic, but they are still support artifacts only. That is live distribution proof, not revenue proof. May 29 measurement truth: the next build was the attribution spine, not more packaging. New files doac/newsletters/newsletter-attribution-baseline-may29.md, doac/newsletters/newsletter-attribution-instrumentation-may29.md, doac/newsletters/may29-naval-proof-capture-pack.md, doac/newsletters/may30-edition-17-link-post-pack.md, and doac/scripts/newsletter-attribution-report.mjs, plus the read-only verifier doac/scripts/verify-newsletter-attribution-source.mjs, now document/report that the live list is still attribution-blind (21 subscribers, 0 source/page/campaign/touchpoint coverage), define the post-deploy verification path, and preserve the next honest proof/distribution steps. Supporting code changes sit across newsletter-worker/src/worker.js, doac/diaryofceo-site/newsletter.html, doac/diaryofceo-site/newsletter-popup.js, and doac/naval-ravikant-diary-of-a-ceo-summary.html; none of that is live proof until worker + site deployment and a test signup verify end-to-end capture. The May 30 follow-on restart surface is doac/newsletters/edition-17-saturday-distribution-launchpad-may30.html, which compresses the next tracked-link/copy/proof sequence into one approval-ready page but still does not count as live traffic, signup, or revenue proof. June 1 truth: doac/newsletters/jun1-attribution-proof-ladder.md narrowed the next honest path to one measured attribution datapoint and surfaced a critical blocker: the deployable newsletter worker path is not yet verified in the workspace, so worker-path confirmation must come before any claimed attribution deployment or proof. June 2 truth: doac/newsletters/jun2-worker-path-tiny-truth-card.md resolved that blocker at file level by proving the canonical worker lives at newsletter-worker/src/worker.js; the next safe attribution step is now one live subscribe-plus-readback check to confirm the deployed worker matches the attribution-aware local source. June 3/4 execution truth: doac/newsletters/jun3-attribution-live-proof-card.md, doac/newsletters/jun3-attribution-smoke-test-runbook.md, and doac/scripts/newsletter-attribution-smoke-test.mjs now define the exact one-command live proof path. doac/newsletters/jun3-overnight-attribution-smoke-test-report.md is the governing verdict artifact right now: the live worker accepts the subscribe and returns the subscriber on readback, but still drops source, formId, pagePath, utm.campaign, and touchpoints. Attribution is therefore a concrete persistence bug lane, not a generic “needs proof” lane. June 5 coverage truth: doac/scripts/audit-newsletter-attribution-coverage-jun5.mjs, doac/newsletters/jun5-newsletter-attribution-coverage-audit.md, and doac/newsletters/jun5-newsletter-attribution-coverage-audit.json now prove the worker bug is only half the lane: deploy source scan found 180 newsletter-capture files with just 2 good, 1 partial, 160 missing attribution, and 17 broken surfaces. That means a worker-only fix would still leave most DOAC signup traffic attribution-blind; the next honest fix order is stale endpoints first, worker-root links second, then the large email-only/plain-form cluster. June 5 daytime execution also produced doac/diaryofceo-site/newsletter-signup-helper.js plus doac/newsletters/jun5-newsletter-attribution-batch-1.md, which implemented the first small repair batch across six pages and created the reusable helper for subsequent batches.
FX Leads DB: fx-leads-landing.pages.dev; blocker is distribution/live buyer conversation, not more packaging. Current buyer-priority stack: fx-leads-db/fx-buyer-shortlist-may6.md, fx-leads-db/fx-top3-buyer-contact-paths-may6.md, fx-leads-db/fx-top3-buyer-outreach-drafts-may6.md, fx-leads-db/fx-top3-buyer-proof-pack-may6.md, fx-leads-db/corpay-pilot-offer-packet-may7.md, fx-leads-db/convera-pilot-offer-packet-may7.md, fx-leads-db/fx-top6-buyer-targets-may7.md, the May 8 send packets fx-leads-db/john-allen-send-pack-may8.md + fx-leads-db/meaghan-riley-send-pack-may8.md, the matching preview/sample assets fx-leads-db/fx-pilot-preview-may8.{html,csv} + fx-leads-db/fx-pilot-sample-50-may8.csv, the May 9 Monday restart assets fx-leads-db/fx-buyer-priority-refresh-may9.md, fx-leads-db/corpay-send-sequence-card-may9.md, and fx-leads-db/fx-pilot-offer-onepager-may9.html, plus the May 10 conversation-support layer fx-leads-db/fx-buyer-fresh-intel-hooks-may10.md, fx-leads-db/fx-first-reply-handling-card-may10.md, and fx-leads-db/fx-pilot-roi-calculator-may10.html, the May 11 touch variants fx-leads-db/fx-monday-buyer-touch-variants-may11.md, the rank-order check fx-leads-db/fx-buyer-path-verification-may11-night.md, the clean #2 reroute pack fx-leads-db/fx-meaghan-riley-routing-pack-may12-night.md, the John-first operator artifact fx-leads-db/fx-john-allen-followup-pack-may13-night.md, the current Friday verdict file fx-leads-db/fx-john-allen-send-decision-pack-may14-night.md, the proof gate fx-leads-db/fx-friday-may15-buyer-touch-proof-card.md, the stricter reroute guardrail fx-leads-db/fx-john-allen-first-touch-reroute-card-may16.md, the proof helper fx-leads-db/john-allen-send-proof-card-may17.md, the no-send-claimed operator note fx-leads-db/john-allen-first-touch-operator-pack-may18.md, the stricter fx-leads-db/john-allen-send-proof-ladder-may19.md, and the May 21 blocker reset fx-leads-db/may21-john-allen-send-gap-card.md. Current send order is now explicitly re-confirmed: John Allen first, Meaghan Riley second, Moneycorp backup third; route AFEX/Cambridge through current Corpay Cross-Border leadership, not stale legacy-brand contacts.
Revenue Products: fastest first-dollar lane is revenue-products/ib-ai-prompt-pack-apr14/. As of May 15, the lane now has a real local sales surface at revenue-products/ib-ai-prompt-pack-apr14/sales-page-may13.html, a grounded bundle in bundle/, AI-Prompts-for-Investment-Bankers-bundle-may13.zip, the exact listing-status memo revenue-products/ib-ai-prompt-pack-apr14/gumroad-listing-pack-may13-night.md, the fallback order page revenue-products/ib-ai-prompt-pack-apr14/direct-sale-order-page-may15.html, the fallback launch memo revenue-products/ib-ai-prompt-pack-apr14/direct-sale-launch-pack-may15.md, and the first warm-demand-test assets warm-outreach-copy-pack-may15-night.md + warm-network-shortlist-may15-night.md, plus warm-demand-test-cockpit-may16.html, warm-demand-test-followup-pack-may16-night.md, warm-demand-test-target-notes-may16-night.md, monday-warm-demand-test-send-sheet-may17.md, warm-demand-test-real-name-candidates-may17-night.md, warm-demand-proof-cockpit-may18.html, warm-demand-real-name-evidence-map-may18-night.md, warm-contact-fill-cockpit-may19.html, warm-demand-real-name-shortlist-may19-night.md, warm-demand-real-name-evidence-pack-may20-night.md, and the new manual-close path first-dollar-manual-demand-runbook-may21.md and the workspace-only blocker proof may22-real-name-hunt-workspace-pass.md. A second May 22 manual-close lane now exists at revenue-products/banker-ai-workflow-audit-may22/ with offer-sheet.md, warm-outreach-pack.md, warm-target-shortlist.md, manual-close-proof-sheet.md, send-proof-log.md, offer-page.html, and intro-route-pass.md; the May 23 follow-ons may23-live-launch-card.md and may23-overnight-real-name-hunt.md compress the lane to its honest blocker state, the May 27 support stack (workflow-audit-roi-calculator-may27.html, workflow-audit-warm-close-console-may27-night.html, warm-route-candidate-grid-may27-night.md) sharpens the close path without changing the grounded-name/send blocker, and the June 7 update added sample-deliverable-jun7.html plus an offer-page.html link to that preview so the warm offer has a tangible sample artifact. Treat it as a higher-ticket warm-network offer that may beat the $19 prompt-pack lane for first dollars, but it is still blocked on grounded warm names, approved payment path, and real send proof. Current blocker truth: the first three prompt-pack warm slots still lack grounded real names in the workspace, so no send-ready fill should be invented. Still no-go until there is a real Gumroad URL/listing, final copy approval, or an explicitly approved manual sales path with an actual live checkout surface.
Funding Intelligence: health must be separated into wrapper reachability, export parity, scrape freshness, enrichment quality, queue hygiene, and upstream write-state. The Jun 12 runs sharpened that rule again. At 7:00 AM, the lane processed 20 pending items and produced 7 enriched / 5 duplicates / 2 non-funding / 6 weak-fit or missing fields, but Tavily failed or was unusable on all 20 lookups and 18 articles fell back to title-only extraction (funding-intel/logs/funding-enrich-run-2026-06-12T11-11-25Z.json). At 12:00 PM, scrape step status degraded to error, direct Python fetch hit Cloudflare 1010, a browser-like curl UA fallback recovered processing, and the batch still only produced 3 enriched / 3 duplicates / 4 non-funding / 10 weak-fit or missing fields while all 16 funding-like candidates fell back to title-only extraction (funding-intel/logs/funding-enrich-run-2026-06-12T16-11-12Z.json). Queue motion happened, but quality was worse, not better. On Jun 14 the lane degraded further into a clean service-state blocker: all three scrape/enrich runs were blocked at the door by 403 Read-only transition Worker: writes and authenticated admin routes are disabled on both /scrape and pending-queue reads, so each run truthfully processed 0 articles and 0 enrichments (funding-intel/logs/funding-enrich-run-2026-06-14T11-00-47Z.json, ...16-00-00Z.json, ...2026-06-15T00-00-00Z.json). Earlier May 27 / May 30 / May 31 degradation history still matters as pattern context, especially title-only fallback and misleading queue-cleanup appearances. The first-stop health surfaces remain funding-intel/funding-truth-dashboard.html, funding-intel/pipeline/last-scrape-enrich-summary.json, and the latest run logs before claiming recovery. A top-level /scrape 200 is not enough to call freshness healthy, synced Pages exports are not enough to call the data current, reduced pending counts are not enough to call enrichment healthy, and a read-only-transition Worker means the lane is hard-blocked upstream until write/admin access is restored. Hermes production plan (May 4) still sets the target architecture: D1 as truth, worker /export/* JSON, no more manual /site/data/ git sync once the rewrite is done. May 19 also surfaced a local news/archive layer inside funding-intel/site/news/ backed by funding-intel/scripts/generate-news-pages.js; keep that product layer inside Funding Intel, not in root dashboards, and do not confuse page generation with freshness/quality recovery.
Blake Engagement Agent: blake-engagement-agent/; prototype lane is now intentionally inactive. The remaining live Blake crons were disabled on May 23 after the sunk-cost call, so treat the repo as archival unless Hunter explicitly revives the project. When Instagram quota is exhausted, preserve last known good post/event state and mark freshness degraded instead of treating the lane as empty. The May 10 1:00 AM and 1:30 AM runs held preserve-state continuity at 189 tracked posts, 273 comment rows, and 21 events under quota pressure, so distinguish that from true fresh-fetch recovery explicitly. On this Mac, run Blake flows through the repo-local py shim (PATH="$PWD:$PATH"; py ... or ./py ...); do not assume a global py binary exists.
MiroFish: /Users/hunta/Projects/MiroFish; graph features still need real ZEP_API_KEY.
Hunter Portfolio: hunter-jackson-ai.pages.dev.
Hermes bot: repaired Apr 29 by moving /Users/hunta/.hermes/config.yaml onto a working openai-codex / gpt-5.5 path and restarting the gateway successfully. Remember the prior failure signatures (gpt-5.2-codex unsupported under ChatGPT-account Codex usage and missing Anthropic credentials) if Hermes regresses.
Kalshi: weather markets, max $10/trade, edge discipline matters more than activity.