Alertmanager does not expand ${ENV} in its YAML, so the committed config with
${TELEGRAM_BOT_TOKEN}/${TELEGRAM_CHAT_ID} crash-looped it (line 24: cannot
unmarshal !!str `${TELEG...` into int64) - 11k+ restarts on prod, alerting dead.
- rename alertmanager.yml -> alertmanager.yml.template (keeps ${} placeholders)
- deploy.sh: envsubst the template into the (gitignored) alertmanager.yml from
.env, scoped to the two TELEGRAM vars so the {{ }} Go-template message survives
- gitignore the rendered file (contains the bot token)
- warns if the vars are unset
Consent gate (the legal linchpin from the wet-signature memo):
- migration 092 adds ink_consent/ink_consent_at/ink_consent_text to esign_records
- extract pure, unit-tested gate logic into esign-ink-consent.ts (DRY single
source for route + signing page): isInkReproduction / inkConsentRequired /
inkConsentSatisfied + verbatim client-safe INK_CONSENT_TEXT
- portal-esign-generic.ts: GET surfaces ink_reproduction + consent text; POST
gates DRAWN signatures on ink-path docs on explicit consent, stores it
- signing page locks the signature block until consent is checked (drawn only)
- npi_provider marks cms855/cms10114 esign metadata ink_reproduction=true
- 33 unit checks: gate truth table + consent text omits all internal mechanics
(plotter/machine/CMS/MAC/etc) and keeps required legal reassurances
Patent-risk memo (docs/legal/patent-risk-mechanical-wet-signature.md):
- prior-art-dated risk analysis (autopen 1803/1942, plotters, CNC = public domain
=> low risk on core concept; e-sign workflow space litigious)
- firsthand recent-grant sweep (1.58M USPTO grants 2021-2025, queried via DuckDB):
ZERO patents on machine-applies-signature-in-ink; e-sign players hold only
electronic-workflow patents. Not an FTO; flags where attorney search is needed
Two internal docs:
- docs/plans/remote-wet-signature-products.md: opportunity map for new remote
signing/filing services that leverage the existing esign + wet-ink + fulfillment
stack (83(b) IRS filings, apostille concierge, estate packages, mechanics
liens, FinCEN BOI / SAM.gov renewals, RON layer, proof-of-life attestations).
Prioritized by revenue x fit x moat; top 3 = 83(b), apostille, estate package.
- docs/legal/remote-mechanical-wet-signature-precedent.md: source-grounded legal
research on whether a machine-applied wet-ink signature (autopen/plotter
reproducing the signer's own captured strokes) is authentic/valid/accepted.
Primary sources retrieved firsthand: DOJ/OLC 2005 autopen opinion (29 Op.
O.L.C. 97); CMS-855B 'signatures must be original'; ESIGN 15 USC 7001/7006;
UCC 1-201 'Signed'. Key finding: common-law + autopen precedent strongly
support own-signature-by-directed-machine as VALID, but 'original ink / no
stamps' administrative rules (CMS-855) are UNADJUDICATED -> highest risk, keep
true wet-sign fallback. Notarized/witnessed instruments: do NOT use plotter.
Explicitly separates established law from interpretive/no-precedent zones.
Adds a second machine class (small fan-shaped reach arm) alongside the
CR-10/AxiDraw rectangular-bed plotters, so wet signatures can be produced
while away from the home station.
ink_signature_plotter.py:
- PlotterConfig gains dialect (marlin|lineus) + name; new LineUsConfig
(native units, pen height = per-move Z, reach annulus from shoulder pivot).
- Named machine profiles (cr10 default, axidraw, lineus) via load_profile().
- bed_mm_to_lineus_units(), check_reach() (annulus for lineus, rectangle for
marlin), compute_jig_offset_for_box() (solves jig from the ACTUAL fitted ink
extent so a wide cell line doesn't over-constrain a small arm).
- emit_gcode() dispatches to emit_marlin_gcode()/emit_lineus_gcode().
- send_lineus(): WiFi TCP 1337 (NUL-terminated, ok-acked) or USB serial,
dry_run=True default (same gating as the CR-10 path).
ink_signature_cli.py: --profile, --solve-jig (auto-applies jig offset),
--lineus-host/--lineus-usb, reach-check that refuses to --plot out-of-reach
on Line-us.
Tests: 43 checks (was 30) covering profiles, reach check, jig solve, lineus
emitter, dry-run sender. Docs updated with profiles + portable workflow.
The admin_todos table is written by 8 worker handlers (and the new shared
create_admin_todo helper) but had NO CREATE migration anywhere, so every
fulfillment-task insert silently failed in environments without the table.
Define it with the exact column set the handlers use, plus status/priority/
order indexes and operator workflow columns (assigned_to, notes, completed_at).
Applied 076,085,086,088,089,090,091 to the dev DB (all idempotent); verified
admin_todos, esign_records, paper_filing_batches, compliance_orders.
fulfillment_status, and esign_records.signature_vector all exist and accept the
handler insert shape.
Telegram notifications:
- Add shared scripts/workers/telegram_notify.py (send_telegram, notify_fulfillment_todo,
create_admin_todo) so every worker alerts the operator the same way; fire-and-forget.
- Fire notify_fulfillment_todo after each admin_todos insert across all 8 service
handlers (9 sites) so no fulfillment task waits unseen.
(Orders + quotes + tickets already notified via checkout/quotes/tickets routes.)
Client portal order progress:
- order-timeline: derive real per-step status from live signals (payment paid,
e-signature signed, fulfillment_status) instead of a static template; add
current_step to the response.
- Extract pure applyLiveStatus into order-timeline-status.ts (DB-free) + unit test
(api/test/test_timeline_status.ts, 8 cases).
- portal /me now returns compliance_orders.fulfillment_status.
- Dashboard renders a client-safe Progress badge (In progress / Action needed /
Filed-awaiting-confirmation / Completed); batches show the most actionable status.
No back-office mechanics exposed.
ERPNext sync parity:
- Create a Sales Order for formation and fcc_carrier_registration orders (previously
only canada_crtc + compliance synced); write erpnext_sales_order back to each table.
Non-blocking, matches existing pattern.
Verified: API tsc clean, timeline unit tests 8/8, Astro build 58 pages,
cms10114/ink/paper_batch Python tests still green, no mechanics leaks.
Adds explicit order timelines for the 5 CMS provider filings (nppes-update,
npi-reactivation, npi-revalidation, medicare-enrollment, provider-compliance-
bundle) and a +2 business-day buffer applied from the signature step onward,
giving us time to produce + mail the original ink signature while the plotter
station is brought online.
- Buffer only affects post-signature steps; pre-signature steps and all
non-wet-signature services are unchanged (verified with date math).
- Single source of truth: order-timeline route (consumed by the order success
page), so the buffer flows through to the customer-facing ETA.
- Remove WET_SIGNATURE_BUFFER_DAYS once the ink-signature station is in
steady-state (see docs/plans/plotter-plan.md).
API tsc clean; buffer compiled into dist.
Source-grounded check of which services need an ORIGINAL ink signature (plotter
target) vs e-sign/typed. Verified firsthand against the official forms:
- Confirmed wet-ink: the 5 CMS Medicare/NPI paper filings only (855I/B/O +
10114), which are exactly the no-login Standard-path filings the plotter serves.
- CLIA CMS-116 does NOT require original ink — the form explicitly permits 'SIGN
IN INK OR USE A SECURE ELECTRONIC SIGNATURE', so our digital stamp suffices;
plotter optional for CLIA.
- DEA registration/renewal is online-only (Form 224 unavailable in PDF),
e-certified, no wet ink.
- State CSR / state Medicaid are the only open items: paper in many states but
original-ink-vs-e-sign is state-specific; verify per state.
- All FCC/telecom/DOT/BOC-3/CRTC/PUC filings are electronic (e-sign fine).
Added the verified matrix to state-healthcare-compliance-opportunities.md, saved
docs/CMS-116 Form.pdf, and the plotter plan.
The Standard no-login CMS path needs an ORIGINAL ink signature on paper
(CMS-10114: 'Stamped, faxed or copied signatures will not be accepted'). This
adds a pipeline to redraw the provider's own captured strokes in real ink with a
pen on a CR-10 V2 (or any Marlin/GRBL machine) — original, in ink, never copied.
- migration 090: esign_records.signature_vector (JSONB stroke paths, 0..1).
- signing page now captures normalized stroke paths alongside the PNG; API
stores a size-bounded vector for drawn signatures.
- ink_signature_plotter.py (hardware-independent): fit strokes to the signature
anchor box, PDF-pt -> bed-mm via jig offset, emit Marlin/GRBL G-code (Z pen or
M280 servo/BLTouch), SVG toolpath preview, and render_signature_on_pdf (a
digital twin that proves the toolpath lands on the cert line). Gated serial
sender (dry_run default).
- ink_signature_cli.py: end-to-end load-record -> gcode+preview, --test-box jig
calibration, --plot to stream over USB.
- Corrected CMS-10114 signature anchor to sit inside the Section 4A signing cell
(above the bottom rule, below the label).
- docs/ink-signature-plotter.md documents the CR-10 retrofit + interpretive risk.
Tests: test_ink_signature.py 30/30, test_cms10114.py 27/27, test_paper_batch.py
15/15, API tsc clean, Astro build 58 pages.
Verified firsthand against the live CMS-10114 (Rev. 02/25, OMB 0938-0931):
- Section 1A confirms paper is valid for Change of Information (#2) AND
Reactivation (#4), not just initial enumeration. Resolves the UNCERTAIN flag.
- Current mailing address is CMS NPI Enumerator Services, Mail Stop DO-01-51,
7500 Security Blvd, Baltimore MD 21244. The old Fargo PO Box 6059 is retired;
corrected in mac_routing.NPI_ENUMERATOR + all docs.
- No electronic no-login equivalent exists for CMS (NPI Registry API is
read-only; PECOS/NPPES-IA require login), unlike FMCSA's ask.fmcsa ticket form.
So tiers stay: Standard=paper CMS-10114 (no login), Expedited=NPPES surrogate.
New: cms10114_pdf_filler.py fills the flat official form via text overlay
(reason checkbox + NPI + Section 2A identity + Section 4A cert name + signature
anchor); wired into npi_provider._generate_10114_for_signing for nppes-update.
Signed forms route to the NPI Enumerator via the existing daily batch.
Tests: test_cms10114.py 27/27, test_paper_batch.py 15/15, Astro build 58 pages.
- NpiIntakeStep: add positively-framed 'can you grant electronic I&A Surrogate
access?' question for all filing slugs (reval/reactivation/nppes-update/
enrollment/bundle). Optional, never required, never mentions paper; captured
as intake_data.surrogate_access (yes/no/blank). Astro build green (58 pages).
- npi_provider.py: surface the surrogate answer in the admin todo so fulfillment
knows EXPEDITED (online via surrogate) vs STANDARD (e-sign + daily mail batch).
Standard (no-login) CMS filings are mailed in one Priority Mail envelope per
destination agency, batched each postal working-day morning to save postage.
- migration 089: paper_filing_batches table + esign_records.paper_batch_id /
filing_destination_key (idempotent: a filing is batched at most once).
- batch_cover_sheet.py: per-agency cover sheet (sender/dest/date/manifest) +
merged print-job PDF (cover + all enclosed signed filings).
- daily_paper_batch.py worker: gather signed+unbatched cms855/cms10114 filings,
group by destination (MAC by state via mac_routing; Fargo for CMS-10114),
build cover+merged PDF per agency, persist batch, mark filings batched.
Self-gates on postal working days (skips weekends + federal/USPS holidays).
Phase 1 = human prints+mails; phase 2 = wire print-mail API.
- worker-crons: pw-paper-batch systemd timer (Mon-Fri 13:30 UTC, self-gated).
- test_paper_batch.py: 15/15 pass (working-day gating, routing, cover+merge).
- Verified Standard(no-login)/Expedited(surrogate) matrix from official CMS-855
PDFs (docs/healthcare-filing-tiers-verified.md): reactivation+revalidation are
855I paper-to-MAC reasons, original-signature, routed by state; sig may not be
delegated; 855B needs PECOS app fee.
- Add scripts/workers/mac_routing.py: state->MAC routing (all 56 jurisdictions,
12 destinations) for envelope addressing + daily batch grouping. Addresses
marked VERIFY before live mail.
- npi_provider.py: fix access strings to two-tier framing; NPPES update/reactivation
no longer 'online-only'; note 855B fee.
- checkout.ts + service pages: strip client-facing mechanics & the paper-vs-tier
choice; surrogate is the only optional, positively-framed ask (faster, never
required, never share password).
