FedM8 automation build spec — Sep 2026
FedM8 automation build — Sep 2026 (CEO-approved 2026-09-03)
Source: hi.jbnx.io v64 review of fedm8.com build board (121/121 done). Repo jbnx/fedm8-scan (gated: PR → test.fedm8.com → chat approve → prod). One claim per PR. Submission always stays human-in-the-loop. Cite-or-blank and read-only Assist guardrails apply to every AI step.
Standing rules for every feature
- After build: verify UX (walk the flow on test.fedm8.com as a Free/Basic/Pro user), functionality (QC spec under
qc/, one live pass), effectiveness (state the user outcome it changes and how it is measured; wire the event into /insights). - Append the feature to
assets/FEATURES.json (done) and log a line in docs/FEATURE_LOG.md with: name, tier, one-sentence customer benefit, compare.html row it strengthens, ad-copy hook. CEO wants this for future compare page + ads. - Verify live keys on the host before asking for any (SAM Entity API, CALC).
Build order (each = one PR)
- SAM-record profile autofill + day-zero saved search — on first sign-in with a UEI/CAGE, pull the public SAM entity record → prefill NAICS, set-asides, POCs; create one saved search + alert automatically. Measure: % new signups with an alert on day 0.
- Cert packet generator — eligibility pre-screen (51% own/control, DD-214/VA letter), ordered document list, one-click ZIP of the user's uploads, self-reported status dates. No submission. Measure: packet downloads.
- Auto-shred on Track — fetch SAM attachments when a notice is tracked, shred without a click, auto-rebuild outline on amendment with a diff notice. Measure: time from Track → shred available.
- SAM renewal + exclusion watch — nightly entity-status diff, 60/30/7 renewal checklist, reps & certs prefill for confirmation. Measure: renewals caught before lapse.
- Firm Library autofill from USAspending — seed past-performance entries from awards the company already holds (by UEI). Measure: library entries per Pro user.
- CALC API pricing — replace links with fetched labor bands in Studio Price. Measure: Price step completion.
- AI red-team pass — Section L/M compliance read before Export; cite-or-blank. Measure: uncited shalls at export.
- Submission package assembler — cover + matrix + drafts + reps + SF-form checklist → one ZIP/PDF plus a drafted CO email. Human sends. Measure: packages assembled.
- Sources Sought auto-response — capability statement draft from Firm Library. Measure: drafts generated.
- Weekly bid-desk digest — extend AI decision brief to a per-user weekly email/inbox item; route Pro win-requests to roadmap. Measure: open/click on
/insights.
Deferred by decision: SLED/eBuy ingest (deep-links stay), any automated VetCert/SAM submission (legal/ToS).