FIN Ad-hoc digest — 2026-07-01_2026-07-02
Across fin-agentic, this window was about turning proof-heavy infrastructure into something operational: the most consequential shift was the Newsroom moving from execution plan and scaffolding into a live editorial dispatch path, complete with renderer, groundedness checks, and coordinator sign-offs. In parallel, the team tightened the RP581 QC authority chain so queue state, provenance, and replay behavior stay trustworthy under load, while comms and Company Brain work kept extending the platform’s intelligence surface.
Highlights
- The Newsroom crossed an important threshold with #1691, which landed brief goldens and the dispatch renderer and marked editorial output as live rather than just planned.
- Supporting that launch, #1690 added groundedness Cats and the first coordinator sign-offs, giving the new editorial flow explicit quality gates and review structure.
- The ad-hoc digest path was wired in via #1688, introducing
makeRangeWindowand range-based output paths without relying on chain/INDEX machinery. - A cluster of QC fixes hardened RP581 as the queue source of truth, with #1687 persisting that authority and #1679 requiring RP581 provenance for queue authority decisions.
- Replay behavior for RP581 extraction also got safer in #1683, which removed queue materialization from the replay path, building on earlier routing and authority-replay fixes.
- Outside the Newsroom/QC push, the platform expanded its intelligence and comms foundation with Company Brain MS Graph wiring in #1667 and merged-conversation identity reads in #1681.
Work started
| Repo | PR | Title | Author | Labels |
|---|---|---|---|---|
fin-platform/fin-agentic | fin-platform/fin-agentic#1693 | newsroom(N-07): Exemplar v2 — real W26 dispatch, engine e2e, Cat-gated 12/12, vault-committed [AWAITING PROOF+OWNER] | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1692 | Company Brain agentic intelligence execution program (draft) | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1691 | newsroom(N-06/N-09 + 4 sign-offs): brief goldens + dispatch renderer — editorial is LIVE | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1690 | newsroom(N-12 + sign-offs): groundedness Cats + first coordinator sign-offs (N-02/N-03/N-08) | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1688 | newsroom(N-08): makeRangeWindow + ad-hoc digest paths (no chain/INDEX) | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1687 | fix(qc): persist RP581 queue source truth | robertolouisgarcia-maker | run-heavy-ci |
fin-platform/fin-agentic | fin-platform/fin-agentic#1686 | newsroom(N0/N1 batch): wire hardening + vault registration + authority bundle + brief scaffolding | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1685 | docs(newsroom): FIN Agentic Newsroom execution program + central proof tracker | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1684 | Comms Phase 1 durable fan-out proof | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1683 | fix(qc): replay RP581 extraction without queue materialization | robertolouisgarcia-maker | run-heavy-ci |
fin-platform/fin-agentic | fin-platform/fin-agentic#1682 | fix(qc): mark load-expected fallback provisional | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1681 | feat(comms): read merged conversation identities | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1680 | fix(ci): accept same-tree proof attestations for deploy | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1679 | fix(qc): require RP581 provenance for queue authority | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1678 | fix(qc): keep RP581 PDF queue rows authoritative | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1677 | docs(proof): record merged master comms phase0 proof | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1676 | Fix RP581 PDF replay tail drain and arrivals queue routing | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1675 | fix(deploy): guard RP581 queue source of truth | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1674 | Persist RP581 PDF authority replay fixes on master | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1673 | Fix RP581 PDF authority replay for QC queue | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1672 | fix(ci): seed staging evals without psql | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1671 | fix(seed): write knowledge embedding version compatibly | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1670 | fix(seed): write knowledge embedding model in tenant seed | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1669 | fix(ci): force eval infra persona seeds | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1668 | feat(qc): harden vision training corpus contract | robertolouisgarcia-maker | — |
fin-platform/fin-agentic | fin-platform/fin-agentic#1667 | Wire Company Brain MS Graph intelligence | robertolouisgarcia-maker | — |
Work completed
Merged pull requests
| Repo | PR | Title | Author |
|---|---|---|---|
fin-platform/fin-agentic | fin-platform/fin-agentic#1691 | newsroom(N-06/N-09 + 4 sign-offs): brief goldens + dispatch renderer — editorial is LIVE | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1690 | newsroom(N-12 + sign-offs): groundedness Cats + first coordinator sign-offs (N-02/N-03/N-08) | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1688 | newsroom(N-08): makeRangeWindow + ad-hoc digest paths (no chain/INDEX) | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1687 | fix(qc): persist RP581 queue source truth | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1686 | newsroom(N0/N1 batch): wire hardening + vault registration + authority bundle + brief scaffolding | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1685 | docs(newsroom): FIN Agentic Newsroom execution program + central proof tracker | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1684 | Comms Phase 1 durable fan-out proof | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1683 | fix(qc): replay RP581 extraction without queue materialization | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1682 | fix(qc): mark load-expected fallback provisional | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1681 | feat(comms): read merged conversation identities | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1680 | fix(ci): accept same-tree proof attestations for deploy | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1679 | fix(qc): require RP581 provenance for queue authority | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1678 | fix(qc): keep RP581 PDF queue rows authoritative | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1677 | docs(proof): record merged master comms phase0 proof | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1676 | Fix RP581 PDF replay tail drain and arrivals queue routing | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1675 | fix(deploy): guard RP581 queue source of truth | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1674 | Persist RP581 PDF authority replay fixes on master | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1672 | fix(ci): seed staging evals without psql | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1671 | fix(seed): write knowledge embedding version compatibly | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1670 | fix(seed): write knowledge embedding model in tenant seed | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1669 | fix(ci): force eval infra persona seeds | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1668 | feat(qc): harden vision training corpus contract | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1667 | Wire Company Brain MS Graph intelligence | robertolouisgarcia-maker |
fin-platform/fin-agentic | fin-platform/fin-agentic#1666 | fix(comms): prove phase0 mobile native detail | robertolouisgarcia-maker |
By the numbers
- PRs opened (work started): 26
- PRs merged (work completed): 24
- Issues closed (work completed): 0
- Commits (supporting signal): 24
- Repositories touched: 1
| Repo | Opened | Merged | Closed | Commits |
|---|---|---|---|---|
fin-platform/fin-agentic | 26 | 24 | 0 | 24 |
Links
- fin-platform/fin-agentic#1693 (opened)
- fin-platform/fin-agentic#1692 (opened)
- fin-platform/fin-agentic#1691 (opened)
- fin-platform/fin-agentic#1690 (opened)
- fin-platform/fin-agentic#1688 (opened)
- fin-platform/fin-agentic#1687 (opened)
- fin-platform/fin-agentic#1686 (opened)
- fin-platform/fin-agentic#1685 (opened)
- fin-platform/fin-agentic#1684 (opened)
- fin-platform/fin-agentic#1683 (opened)
- fin-platform/fin-agentic#1682 (opened)
- fin-platform/fin-agentic#1681 (opened)
- fin-platform/fin-agentic#1680 (opened)
- fin-platform/fin-agentic#1679 (opened)
- fin-platform/fin-agentic#1678 (opened)
- fin-platform/fin-agentic#1677 (opened)
- fin-platform/fin-agentic#1676 (opened)
- fin-platform/fin-agentic#1675 (opened)
- fin-platform/fin-agentic#1674 (opened)
- fin-platform/fin-agentic#1673 (opened)
- fin-platform/fin-agentic#1672 (opened)
- fin-platform/fin-agentic#1671 (opened)
- fin-platform/fin-agentic#1670 (opened)
- fin-platform/fin-agentic#1669 (opened)
- fin-platform/fin-agentic#1668 (opened)
- fin-platform/fin-agentic#1667 (opened)
- fin-platform/fin-agentic#1666 (merged)
Provenance
Auto-generated by the scripts/maint/generate-gh-digest@1.0.0 projector from gh query output.
- Owner scope:
fin-platform(all active FIN repos) - Window:
2026-07-01T00:00:00Z→2026-07-03T00:00:00Z - Sources:
github:prs.opened,github:prs.merged,github:issues.closed,github:commits - Lede + Highlights: LLM editor (
LLMProvider) - Regenerate:
pnpm digest:gh --from 2026-07-01 --to 2026-07-02