Host-composed command replies, runtime and admission diagnostics, and
lifecycle notices now carry role="system" plus a descriptive system_type at
their producer. send_with_budget persists both fields on progress records as
well as ordinary chat rows, and gateway/history.py replays the recorded voice,
so a reload no longer re-narrates host text in Ouroboros's own voice. Rows
written before these fields existed keep the assistant default.
Model speech stays model-authored: round narration, proactive replies, final
answers and question pointers are selected by the existing narration fact
rather than by reading message text.
The Host Service named-operation view accepted only direction="out" terminal
rows, so a typed command reply would have stayed pending forever; it now
recognizes either output voice while keeping the exact-origin ownership check.
tests/test_host_message_voice.py scans runtime send calls and literal chat
envelopes with counted, reasoned model/transport exceptions, so a new
unstamped producer or a stale exception fails the suite; computed aliases stay
a documented review duty. web/tests/host_message_voice.test.js pins the real
Chat consumer: System versus assistant voice, escaped System text, a typed
nonterminal notice that must not finish a working card, and an explicit
terminal fact that still does.
DESIGN defines chat authorship, the System row and the markdown asymmetry;
CHECKLISTS item 30 now covers backend UI-message producers, which closes the
stale pointer in the design-system chapter. The architecture chapter's byte
budget is raised because the previous raise consumed its own headroom and this
description replaces nothing.
Out of scope and unchanged: the message-bus role default, the history replay
gate, PROJECT_ROW_TYPES, question cards, the untyped terminal-host-notice
contract (#1010), the palette-token contradictions (#1125), and the
pre-existing clearing of live-header state by unkeyed notices.
Co-authored-by: Ouroboros <311266734+ouroboros-agent@users.noreply.github.com>
All 15 D15 runtime modules are unsplit by design - upstream bytes stand as-is:
7 byte-identical to the reference, 6 pure upstream drift (nothing to do), and
2 (consciousness, reflection) carry v7 deltas of the D02 family that must NOT
be replayed verbatim - reflection's status-set delta would invert over
upstream's own handling (the re-prove trap is documented in
docs/v7next/LEDGER_CORRECTIONS.md together with a MIGRATION row already
superseded by upstream).
The test side transplants the oracle's D15 split: the 2386-line evolution
integrity giant becomes six themed suites (lossless: 65==65 test functions,
zero upstream drift of the giant since merge-base) plus three siblings carried
verbatim; adaptations are exclusively reverse-mappings of OTHER domains' v7
spellings back to upstream signatures, each keyed to the original monolith.
98 passed in isolation (re-verified independently); every file <=617 lines;
HEAD held through six pytest runs.
