Forrige commit skrev en CLAUDE.md-invariant om at en betalt test far sin egen
opt-in, men gatet bare den NYE testen. test_portfolio_live.py passerer ingen
client_factory og er derfor selv en betalt kjoring - den fyrte pa
to-variabel-paret fra et bart `uv run pytest`. Invarianten var altsa halvt
usann den dagen den ble skrevet, som er nyaktig Fase 3-klassen (en pastand
flaten gjor om seg selv uten dekning).
MALT etter fiksen, med begge Foundry-vars satt og UTEN PORTFOLIO_LIVE_FULL_RUN:
test_foundry_profile_live PASSED (det billige trinnet, med vilje pa to vars)
test_portfolio_live SKIPPED (var betalt og ugatet for dette)
test_full_run_live SKIPPED
Ovrige oppfolgingspunkter fra review:
- assert_full_run_contract far ekte type hints (repoets "type hints overalt";
mypy src ser ikke tests/, sa den ville rotnet ubemerket).
- Maleprotokollen SS6 far oppsettet STATE peker pa - model-map-formen og de fire
export-linjene. Pekeren lovet noe seksjonen ikke hadde.
869 passed / 5 skipped, uendret. ruff+format+mypy rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GF7va4cpRiuf79kTzAi3vW
Fase 1b siste trinn: forste bundne levende kjoring over HELE run_project-stien
mot Foundry (gpt-4-1-mini). Okt 37s uttalte aerlighets-grense er lukket ved
maling: ingen -parse-failures.json i outboksen, altsa parset hvert eneste
genererings-svar. assumptions-normaliseringen virket ende-til-ende.
Utfall: rejected pa P90 (claimed 34500 > feasible 11488), checker approve.
Kjoringen KONKLUDERTE - per pre-registreringen det bestatte utfallet. De to
falsifisererne skilte lag for forste gang mot en levende modell.
FUNN storre enn den gronne testen: modellen fant opp kostkoden
EL-LIGHTING-OP-HR (null treff i kunnskapsbasen). Avvisningen var riktig men
skjedde pa 30%-cap-en, ikke stage 0 - bundelen shipper ingen cost-baseline.json,
sa S4.0-forankringen var inaktiv. Ko-fort, ikke rettet her.
Gate-designet er ovis beslutning, tatt for koding:
- TREDJE distinkt opt-in PORTFOLIO_LIVE_FULL_RUN (truthiness, 4b-invarianten).
Begge eksisterende live-tester gater pa SAMME to Foundry-variabler, sa
gjenbruk ville latt den billige proben fyre den dyre kjoringen - stigen i
maleprotokollen ville kollapset til ett trinn. MALT: den dyre SKIPPET med
begge Foundry-variablene satt.
- Asserten i EN kopi (conftest.assert_full_run_contract, ko-(p)), smal med
vilje: fravaer av parse-failures-artefaktet + at validatoren avgjorde. En
rejected BESTAR - pastanden er schema-aksept, ikke modell-dommekraft.
- Iron Law uten a betale to ganger: diskrimineringen bevist OFFLINE av
test_live_full_run_contract.py. To mutasjoner, hver sin signatur: detach
artefakt-sjekken (T1 rod ALENE) - raise ubetinget (T2 rod ALENE).
STATE-premiss korrigert: "test_foundry_profile_live dekker KUN klient-nivaet"
var upresist - test_portfolio_live.py dekket allerede fan-outen, men dens
len(runs)==1 kan ikke skille validert fra avvist og bar derfor ikke pastanden.
869 passed / 5 skipped (fra 867/4), ruff+format+mypy rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GF7va4cpRiuf79kTzAi3vW
Prosjektets første levende kjøring døde på `rounds limit=12 observed=13`, og den
hostede flaten svarte `500 {error_type, error}` — nøyaktig det samme den sier når
modell-endepunktet faller. Nå: 429 med trippelen som STRUKTUR.
Beslutningen er S3.4-invarianten anvendt på transporten: `budget_stop` ble holdt
utenfor `stop_reason` fordi de to stoppene betyr motsatte ting, og å svare
ressurs-utmattelse på krasj-kanalen gjør «det gikk ikke» uleselig på samme måte.
IKKE 200, og det er dét som skiller den fra `Rejection`: en `Rejection` er en
kjøring som KONKLUDERTE og hører i payloaden, mens et uttømt budsjett produserte
ingen proposal — en 2xx ville latt en automatisk kaller bokføre «analysert» for en
kjøring som analyserte ingenting. 429 fordi betingelsen oppstår av en TILDELING
(`max_rounds`/`max_tokens` er whitelistede request-felt), aldri av en serverfeil.
`kind`/`limit`/`observed` legges ut som felt, aldri `str(exc)` (kø-(y));
`error_type` holdes ute — den nøkkelen tilhører feilkanalen. `budget_exhausted` er
ikke foldet inn i `outcome_type` og kunne ikke vært det: `outbox.outcome_payload`
er den ene kopien av den forgreningen og tar `ValidatedProposal | Rejection`.
Ærlighets-grense: ingen `Retry-After` — å vente endrer ingenting.
Iron Law: begge nye tester RØDE før armen fantes. Fem mutasjoner mot HELE suiten,
alle røde med hver sin signatur, grønn kontroll 867/4: detach armen (2 røde) ·
flat streng i stedet for struktur (1 rød — struktur-testen alene) · ekko `limit`
som `observed` (1 rød) · utvid armen til `Exception` (6 røde) · stemple
`error_type` på budsjett-kroppen (1 rød).
500-armens vitne ble BYTTET, ikke slettet: den eksisterende testen brukte
`BudgetExceeded` som sin 500-prøve, så en ny arm alene ville etterlatt
krasj-kanalen uten vitne. Den bærer nå en ekte ikke-budsjett-`RuntimeError`, og er
dét som holder den nye armen smal.
Kjørt, ikke bare testet: `python main.py` startet, `/readiness` 200, ukjent felt →
400 med navnet, 404, SIGTERM → exit 0.
865 → 867 passed / 4 skipped; ruff + format + mypy rene. DEPLOY.md §6 dokumenterer
429 for mottakeren.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W1bsX79aDS7fJ5udWGWAEN
Operatørdirektiv 14.08 etter ekstern test: mottakeren skal få kjørbar Python, ikke en
Docker-innpakking. `Dockerfile` og `azure.yaml` er SLETTET fra treet.
Sømmen er valgt av den eksisterende invarianten, ikke av smak: pakka ER `git archive HEAD`,
så å ekskludere filene fra arkivet ville krevd en kurerings-mekanisme — den andre kopien av
«hva mottakeren får», fri til å drifte fra HEAD (kø-(p)). Fjerning holder arkivet ukurert og
gjør fraværet til en egenskap ved HEAD, som er det eneste en gate kan måle.
De to gatene som pinnet flaten er håndtert bevisst:
- 4e-rå-tekst-gaten (`--platform linux/amd64` + ÉN kopi av startkommandoen) er SLETTET, med
et notat der den sto. En gate som pinner en fjernet flate kan bare bli grønn.
- handover-gatens `_REQUIRED_MEMBERS` er ikke bare fratatt de to navnene, men erstattet av en
POSITIV fraværs-assert pluss en dokument-gate. Å kun slutte å KREVE dem ville gitt en gate
som ikke kan skille «fjernet» fra «shippes fortsatt».
Startkommandoen har nå ÉN kopi igjen: DEPLOY.md-ens `python main.py`, som navngir inngangen
subprosess-testen faktisk kjører.
Fase 1b, funn 1b. Den første levende kjøringen brant tolv runder på svar som
ikke lot seg parse til IR-formen; e371890 gjorde teksten synlig, dette fjerner
årsaken. generate_via_llm sender nå
options={"response_format": proposal_response_format()} på hvert
genererings-kall.
Formen er MÅLT, ikke valgt. ChatOptions.response_format tar
type[BaseModel] | Mapping, og begge profiler ærer den: LOCAL sender en Mapping
ordrett til Chat Completions, AZURE (FoundryChatClient -> RawFoundryChatClient
-> RawOpenAIChatClient) konverterer samme envelope til Responses-APIets
text.format. Klassen — det korteste svaret — er avvist på bevis: gitt en klasse
konverterer klienten med type_to_response_format_param, som emitterer
minimum/exclusiveMinimum/minItems/prefixItems og et assumptions-node hvis
additionalProperties er et skjema. Azures publiserte subset utelukker alle fire.
assumptions kan ikke bare droppes, og det er også en måling: validator
._monte_carlo faller tilbake på item.unit_cost for hver kode uten bånd, så uten
bånd er alle 512 samples identiske og P10 == P50 == P90. Den stokastiske
falsifisereren ville gått inert mens den fortsatt rapporterte persentiler.
Wire-en bærer derfor et array av navngitte entries som _parse_ir folder tilbake
til IR-ens map — additivt, aldri erstatning. Skjemaet deriveres fra
SavingsProposal; sanitiseren er fail-closed (StructuredOutputUnsupported).
Load-bearing målt mot hele suiten, seks mutasjoner alle røde, grønn kontroll
864/4: detach wiringen (1) · detach sanitiseren (3) · dropp assumptions fra
skjemaet (1) · fail-closed -> stille reparasjon (1) · detach normaliseringen
(3) · erstatning i stedet for tillegg (2, inkl. golden-transkriptet).
T3 ble skrevet vakuøs først og felt av sin egen mutasjon: den påsto å bli rød
når assumptions forsvant fra skjemaet, men den scriptede klienten ignorerer
skjemaet. Testen fikk en direkte assert på skjemaet.
Ærlighets-grense: ingen betalt kjøring gjort. Testene beviser konformitet med
det dokumenterte subsettet, ikke aksept fra det levende endepunktet.
859 -> 864 passed / 4 skipped. ruff + format + mypy rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013EQNU4tfAhsBvdefT1jUhk
Fase 1b funn 1 lukket. generate._fetch_parsed kastet hvert uparsebart modellsvar i
except: continue, så prosjektets første levende kjøring brant tolv runder på formatfeil
og etterlot null tegn av det modellen faktisk sa. Enhver videre betalt kjøring ville
vært gjetning.
HVOR teksten overflates er avgjort av en måling, ikke av symmetri med Steg 5:
meter.tick_round() raiser BudgetExceeded INNE i _fetch_parsed, og uten mandat fanger
ingen den, så på nøyaktig den stien fangsten finnes for returnerer generate_via_llm
ingenting. Et felt på GenerationResult ville vært blindt for den, og et outbox-artefakt
skrevet etter kjøringen likeså. Sinken speiler meter: en kaller-eid akkumulator hvis
innhold kalleren holder uansett hvordan løkka endte. Artefaktet skrives fra en finally,
ikke except BudgetExceeded, og kun når noe faktisk feilet.
Iron Law: testfila rød ved collection FØR modulen fantes. Seks mutasjoner mot HELE
suiten, alle røde, hver med sin egen signatur; grønn kontroll 859 passed / 4 skipped
(fra 854). Den skarpeste er trunkering som BEHOLDER sentinelen: da faller kun
verbatim-asserten, som er det som beviser at den ene testen bærer den egenskapen.
Samme økt: mutasjonsmålingen økt 34 utsatte for de to Fase 5-gatene er kjørt. Fire
preflight-mutasjoner mot hele suiten, alle røde på riktig test og ingen annen (detach
fallbacken 2 røde, snu presedensen, presence i stedet for truthiness, avslaget navngir
kun vårt navn). To handover-mutasjoner kjørt MÅLRETTET mot egen testfil under tidspress,
ikke mot hele suiten — uttalt, ikke skjult (drop uv.lock, bygg fra arbeidstreet i stedet
for tracked files). De to DEPLOY.md-mutasjonene gjenstår: git archive leser HEAD, ikke
arbeidstreet, så de krever en midlertidig commit.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WLQd8ojQ9xwxhB8vsETYBs
Målt fra den utpakkede overleveringspakka: med KUN plattformens injiserte
FOUNDRY_PROJECT_ENDPOINT — altså nøyaktig situasjonen i en hostet Foundry-container —
avslo preflight en konfigurasjon backends.py ville godtatt. Gaten og kjørestien kjente
ulike navn; det er repoets egen «checker og kjøresti validerer ulikt»-klasse, og for
mottakeren av pakka er det et falskt avslag på riktig oppsett.
_ENDPOINT_ENVS IMPORTERES nå fra backends i stedet for å gjentas, så de to kan ikke
drifte fra hverandre igjen. Presedens over VERDIER, ikke deklarasjoner: et eksportert-men-
tomt eget navn faller igjennom i stedet for å skygge et ekte injisert inn i en fail-fast.
Avslaget navngir BEGGE variablene.
Iron Law: 3 røde diskriminatorer + 1 grønn kontroll FØR fiksen. 854 passed / 4 skipped.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SeW1LhH5TtXxKZPe9JkqL1
Én zip en mottakende organisasjon deployer uten å klone repoet eller ha konto her.
Arkivet er git archive HEAD (tracked files only), som er SAMME tre den målte
docker-build-konteksten bruker — og grunnen til at STATE.md/*.local.md/.env ikke kan
komme inn: de er gitignorert, ikke filtrert bort av et filter vi må vedlikeholde.
DEPLOY.md svarer mottakerens tre første spørsmål: hvem gjør hva (plattform-operatør,
bestiller, fagperson), prosessen ende-til-ende, og hvorfor det ikke finnes et
chat-grensesnitt. Den navngir også deploy-kravet 4e målte men aldri skrev ned:
pakket model_map.json bærer REPLACE-WITH-*, så uten PORTFOLIO_MODEL_MAP starter
containeren, svarer på /readiness og feiler hver invocation.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SeW1LhH5TtXxKZPe9JkqL1
Funnet av ren-klon-verifikasjonen, og KUN der: to modul-docstrings bærer ``\```
(en escapet backtick i prosa om at rå ``|`` og ``\`` overlever uendret gjennom
http-sømmen). Python advarer om ugyldig escape-sekvens ved KOMPILERING, så i
arbeidstreet er advarselen borte etter første kjøring — .pyc-cachen gjør den
usynlig for oss og synlig for alle andre. Nøyaktig den klassen denne fasen
finnes for: en defekt som bare viser seg på fremmed maskin.
Fikset ved å gjøre de to docstringene rå (r"""). Innholdet er BYTE-UENDRET —
prosaen som beskriver escapingen skal fortsatt lese som den gjør; det er
tolkningen av den som var feil, ikke teksten.
Målt: 0 SyntaxWarning igjen ved kompilering av alle .py utenom .venv (var 2).
846 passed / 4 skipped, ruff + format rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011ckyg3Pc6k7FRuR6fDGQLJ
Fase 3 (AAA+ på publisert flate). Tre av planens premisser falt på måling og er
rettet FØR handling, ikke etterpå:
* GOVERNANCE-raden hadde feil tiltak. Planen sa «skriv den»; org-ops D11 sier én
kanonisk fil som hvert repo LENKER, og filen er nå publisert (målt: HTTP 200 på
open/repo-standard). Å skrive vår egen ville gjort oss til kopi nr. 12 av en
fil D11-bølgen holder på å rydde vekk. README lenker den, i samme form som
repo-mailbox bruker, og bus-faktor 1 står uttalt i den kanoniske teksten.
* Release-objektet for v1.0.0 FINNES allerede på open/ (id 155, CHANGELOG-kropp,
siden rendrer) — det som mangler er vedlegg, ikke objektet.
* WARN RELEASE-STALE fyrer ikke, og kan ikke: regelen sammenligner utgivelse mot
tagg og er strukturelt blind for repo med null utgivelser (org-ops hovedbok
#18). Gaten var OK/20 sjekker FØR arbeidet startet, så den kan ikke tjene som
verifikasjon for denne fasen. Bevisene er Forgejo-APIet, filinnholdet og
ren-klon-kjøringen.
A5-defekten rettet: env.template:21 sa at credential resolves via
DefaultAzureCredential. Den har aldri gjort det — backends.py:149 konstruerer
ManagedIdentityCredential eller AzureCliCredential, og Learns MAF-veiledning
navngir den spesifikke credentialen NETTOPP for å unngå probing. En operatør som
kopierte templaten ble fortalt at feil identitet ville bli brukt.
To load-bearing gater (Iron Law: begge røde før fiksen, 2 failed / 7 passed):
1. env.template navngir de credentials backends.py faktisk konstruerer, og ingen
linje utgir DefaultAzureCredential for å være mekanismen. LINJEFORANKRET, ikke
delstreng: backends.py NAVNGIR klassen fire ganger i kommentarene som
begrunner hvorfor den ikke brukes, så en fil-bred substring-gate ville vært
rød på nøyaktig den prosaen den beskytter (repoets 08-09-klasse, fjerde gang).
2. README-ens wheel-filnavn bærer versjonen bygget stempler på fila. Uten den
ville en versjonsbump stille etterlatt en publisert install-kommando som peker
på en fil som ikke finnes.
Hver positiv assert er paret med en KONTROLL på at det søkes etter noe som
finnes — en ekstraktor som stille finner null lager en gate som bare kan bli
grønn.
MUTASJONER MÅLT MOT HELE SUITEN, begge røde på riktig test og på INGEN annen:
gjeninnfør den usanne credential-påstanden (2 røde, 844 grønne) · la
wheel-filnavnet drifte til 1.0.0 (1 rød, 845 grønne). Restaurert fra scratchpad
+ shasum -c mellom hver. Bumpen selv var den andre mutasjonen: pyproject 1.0.0 →
1.1.0 gjorde README-gaten rød alene, før README ble rettet.
SECURITY.md: varslingsfrist (minst én minor-release og aldri under 30 dager
mellom kunngjøring og fjerning, med sikkerhetskritisk fjerning som uttalt
unntak). Støttetabellen er bevisst VERSJONSFRI — et release-nummer skrevet der
ville drevet ved neste tagg, altså samme defektklasse som gate 2 fanger.
CLAUDE.md beholdt på flaten med en engelsk innramming øverst (operatørvalg): den
sier hva fila er for en fremmed. Innholdet er repoets sterkeste bevis på at hver
beslutning er målt; å fjerne det ville fjernet bevis, ikke friksjon.
Versjon 1.1.0 — synket i pyproject, __init__, test_smoke og README-kommandoen.
1.0.0-treet kan ikke produsere en kjørbar wheel (force-include kom etter taggen,
målt: git show v1.0.0:pyproject.toml har den ikke), så en wheel hengt på den
utgivelsen ville vært nøyaktig den usanne påstanden denne fasen finnes for å
fjerne. Operatøren valgte bumpen framfor et vedlegg som ikke virker.
846 passed / 4 skipped (fra 837). ruff + format + mypy rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011ckyg3Pc6k7FRuR6fDGQLJ
De to gjenstående 4e-radene, begge målt mot hele suiten (837 passed / 4 skipped).
(1) /invocations svarer gyldig mot en SKRIPTET backend gjennom EKTE run_project.
Alle 4d-testene ga invoke en stand-in som sluker **kwargs, så whitelisten kunne
navngi et felt run_project ikke tar — eller sende samme argument to ganger — uten
at én test merket det, mens en levende container svarte 500. Sømmen er
run._default_factory, ikke payloaden: client_factory nektes av whitelisten med
vilje, så factory-defaulten er eneste injeksjonspunkt flaten etterlater.
Payloaden sender HVERT whitelistet felt, med en dekningsassert mot
_ALLOWED_FIELDS. Profilen er LOCAL fordi AZURE-armen slår opp et
Foundry-deployment-navn i modell-mappet FØR noen klient bygges (målt).
(2) Rå-tekst-gate: Dockerfile + azure.yaml kjøres av ingen test (docker build og
azd deploy er operatør-gatet). Gaten pinner --platform linux/amd64 (målt påkrevd)
og ÉN kopi av startkommandoen (imagets CMD; azure.yaml har ingen startupCommand).
Nøkkel-sjekkene er linjeforankret, ikke delstreng — azure.yaml sin egen kommentar
navngir begge nøklene for å begrunne fraværet.
Fem mutasjoner, alle røde på riktig test og på INGEN annen (836 øvrige grønne
hver gang): send project_id to ganger · whitelist et felt run_project ikke tar ·
fjern bundle_dir fra whitelisten · fjern --platform linux/amd64 · gi azure.yaml en
startupCommand-nøkkel. Kontroll: pristine tre 837/4.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018GfbDLY7YVLKVqpUHnbwVW
To målinger avgjorde formen FØR koden: (1) hosting-pakkas InvocationsHostServer
finnes kun i bygg som krever agent-framework-core>=1.13.0 (treet låser 1.9.0;
eneste 1.9-kompatible bygg er en forlatt alfa som importerer mcp udeklarert),
(2) et gjenbrukt bygget workflow er single-use på 1.9.0 (kall-serie [2,0,0] —
rundetaket persisterer; ferskt objekt per kall er ren kontroll). Derfor spikens
§5-fallback: hosting.py serverer kontrakten (8088/PORT, /readiness,
/invocations, SIGTERM→0) selv, stdlib asyncio på ÉN løkke — aldri as_agent()
(gatene ligger utenfor grafen), aldri tråder (NG1-guarden fanget første utkast
med ThreadingHTTPServer; asyncio-formen består den by construction).
Payload whitelistes på run_projects signatur — ukjente felt nektes ved navn
(400), aldri stille droppet; profile defaulter til azure kun her. ValueError →
400, alt annet → 500 {error_type, error}; Rejection er vellykket kjøring → 200.
outbox.outcome_payload ekstrahert som den ENE kopien av validated/rejected-
forgreningen (kø-(p)-regelen). azure.yaml validert GRØNN mot begge autoritative
skjemaer (jsonschema, hentet ferskt); ingen env:, ingen startupCommand (imagets
CMD er den ene kopien). Dockerfile: 3.12-slim-bookworm + git + uv==0.9.8 +
uv sync --frozen --no-dev; git archive <indeks-tre> | docker build
--platform linux/amd64 grønn på nøyaktig de stagede bytene.
Iron Law fulgt: testfila rød ved collection FØR modulen fantes. 835 passed /
4 skipped (fra 821), ruff+format+mypy rene. Seks mutasjoner mot HELE suiten,
alle røde på riktig test: detach felt-mappingen · dropp ukjente felt stille ·
flipp 400/500 · detach azure-defaulten · detach SIGTERM-handleren · detach
main.py-shimen (de to siste kun fanget av subprosess-testen, P4-presedensen).
Deploy IKKE utført — azd-steget er operatørens.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PEiiSGRShizKc771ZBa1iq
Endepunktet løses som første ikke-tomme av vårt eget
PORTFOLIO_FOUNDRY_PROJECT_ENDPOINT og Foundrys injiserte
FOUNDRY_PROJECT_ENDPOINT — vårt vinner, fallbacken lar samme image kjøre
hostet uten ekstra wiring. Presedensen gjelder verdier, ikke deklarasjoner.
Credential velges av samme miljø: AzureCliCredential lokalt,
ManagedIdentityCredential når FOUNDRY_HOSTING_ENVIRONMENT er satt, fordi
containeren ikke har noen Azure CLI. Ikke DefaultAzureCredential — Learns
MAF-veiledning navngir den spesifikke credentialen for å unngå probing.
Load-bearing målt mot hele suiten, fire mutasjoner alle røde + grønn
kontroll: detach credential-valget · presence i stedet for truthiness ·
detach fallbacken · snu presedensen. Fail-fast-testen var vakuøs først —
vårt variabelnavn inneholder det injiserte som delstreng.
De fire åpne azure.yaml-valgene lukket mot de to JSON-skjemaene og ført i
docs/2026-08-13-fase4-azure-yaml-valg.md. Ingen azure.yaml skrevet (4d).
821 passed / 4 skipped. Ruff + format + mypy rene.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Jk8tauXXAojNKC7Tzq7ziF
README:24 sa det rett ut: shared/ ble lest fra arbeidstreet ved kjøretid, og
derfor kunne repoet verken publiseres som wheel eller kjøre i container. Målt
før endringen: 1.0.0-wheelen bar 58 filer, null under shared/.
Endringen er én søm + én pakkelinje:
- hatchling force-include speiler shared/ byte-identisk til
portfolio_optimiser/_shared/ (wheel 122 filer, 64 under _shared/; sdist
bærer treet, målt via uv build sdist→wheel)
- shared_root() løser ved kall-tid: PORTFOLIO_SHARED_ROOT → arbeidstreets
shared/ når det finnes (en checkout er autoritativ — det holder pull-only-
subtree-kontrakten og goldenene urørt) → pakket kopi
Iron Law fulgt: tests/test_shared_packaged_data_loadbearing.py skrevet FØRST,
alle tre røde mot dagens kode (ordnings-testen felt av sin egen kontroll på at
pakket kopi finnes). Deretter fiks, deretter MÅLT mutasjon mot hele suiten:
- detach fallbacken → 1 rød (resolusjons-testen)
- detach force-include → 3 røde
- snu rekkefølgen (pakket før arbeidstre) → 1 rød (ordnings-testen, som var
grønn før fiksen — flip-mutasjonen er beviset på at den diskriminerer)
Kontroll grønn: 813 passed / 4 skipped (baseline 810/4 målt på 142bfa9 samme
økt). Goldenene byte-uendret før og etter (shasum -c på demo-transkript +
begge nav-goldens). shared/ selv er urørt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018hfm6sWTk17Cbh6ZHYhvCu
Subtree-pullen (71b7b66) landet commons' domene-nøytrale formulering i
persona-artefaktets `rationale` («i kontorbygg» → «i tilsvarende anlegg»;
commons 73136eb — én fil, én linje, verifisert av OSS mot commons/main FØR
pullen). Linja printes ordrett i demoens Steg 7, så det pinnede transkriptet
sto rødt i pullens egen commit. Mellomtilstanden er uunngåelig: squash-commiten
kan ikke amendes (den bærer git-subtree-trailerne pullen leser neste gang).
Regenereringen er en RE-MÅLING, ikke en kopiering (testens egen docstring:
«whoever regenerates re-takes those measurements»). Den nye linja ble PREDIKERT
fra det innkommende artefaktet FØR pullen, og er byte-identisk med det
programmet faktisk skrev. Prediksjonen var bærende: `_clip(..., 300)` kapper på
fast bredde, og byttet er +8 tegn, så HALEN flytter seg også («overestimerte …»
→ «overes…»). Uten en nedskrevet prediksjon kan en regenerering ikke skille den
forventede forskyvningen fra drift — den ville bare vært «kopier det demoen
skriver nå».
MÅLT:
- rød-settet var NØYAKTIG én test (809 passed / 1 failed / 4 skipped) — ingen
annen assertion matcher en delstreng av `rationale`; commons advarte om
nøyaktig den klassen, og prediksjonen var at kun goldenen ville falle
- gammel vs ny fasit: KUN linje 32
- K1 distinkt 8 (linjer 9 — kjent, dokumentert avvik) · K2 hypotese #1 REJECTED
2 100 000 mot P90 1 769 915 → VALIDATED 445 500, samme kandidat · K6
selv-identitet: to kjøringer byte-identisk stdout, eneste stderr-diff er
po-sim-suffikset · K8 bygg-energi-mikro + nav-golden-* uendret · 61/4 linjer
- stderr-goldenen URØRT · `pyproject.toml` og `uv.lock` urørt av pullen, så
fersk-klon-målingen (P4 pkt. 1) står — delta er prosa i `shared/` + fasit
- 810 passed / 4 skipped · ruff check + ruff format + mypy rene
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017og6HMcP1WQcABRDogUMfx
Version 1.0.0 across the four sites that carry it — pyproject.toml, __init__.py,
uv.lock, test_smoke.py. Measured that these are the only four: README carries no
version badge, and CHANGELOG's `## [0.1.0]` is history rather than a bump site.
The heading stays `[Unreleased]`. STATE authorises the CHANGELOG CONTENT now and
holds the TAG until after Wednesday's freeze, so stamping `## [1.0.0] - <date>`
today would be a future-dated claim about an event that has not happened — and
one to rewrite if the dress rehearsal fails or the freeze slips. `pyproject` at
1.0.0 with `[Unreleased]` populated is the release-prep state, not an
inconsistency; nothing machine-reads the CHANGELOG (measured). The global
versjonssync rule is read as CONTENT, not heading. Tag day is then one atomic
move: rename the heading, stamp the date, tag.
The re-lock was the hazard, and it was gated rather than assumed. Bumping the
version stales `uv.lock`, and the next `uv run` would have re-locked it
invisibly against a RANGE dependency (`agent-framework-core>=1.9.0,<2`) — while
the two ExperimentalWarning lines are pinned byte-for-byte in the stderr golden,
and STATE's own okf note records that a bare sync is enough for a guard to stop
guarding with no local diff. So: bump, then `uv lock` EXPLICITLY, then diff
before any test ran. The diff is the single `portfolio-optimiser` version line;
agent-framework-core, llm-ingestion-okf (v0.3.2) and llm-ingestion-guard
(v0.3.4) are untouched, and uv.lock was re-checked AFTER the suite to confirm no
silent re-lock.
CHANGELOG prose for the six feat commits `[Unreleased]` did not cover — it
carried only Step 5 and the scripted registry. Console entry points and the
golden transcript are Added; the Step-7 inbox, the anchored walkthrough, the
stderr damping and the derived provenance sentence are Changed, scoped as the
OFFLINE SIMULATION rather than framework runtime, since they change what the
walkthrough exercises and not the library's behaviour. The content gate is
Security, and carries its opt-in qualifier: `materialize` stays ungated by
design and `materialize_gated` is asked for by name — an entry claiming "ingest
now scans content before writing" without that clause would overclaim, and it
sits next to the sentence read on stage Thursday. A Notes line names the two
open boundaries (ingest stamp spec divergence, D7 mirroring) so 1.0.0 reads as a
stable surface rather than a finished programme.
Measured, not asserted: 810 passed / 4 skipped unchanged · ruff + mypy clean (31
source files) · no `0.1.0` remaining outside .venv/shared · and the demo RUN, not
just tested — stdout byte-identical to tests/golden/demo-transcript.stdout, exit
0, 61 stdout / 4 stderr lines, matching dress rehearsal nr. 0. The version string
appears nowhere in either golden (0 hits), so the bump could not move the fasit.
Two STATE premises corrected by measurement: 24 commits since v0.1.0, not 23;
and eight feat commits exist since the tag, of which six were undocumented.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ue1AnPZYsC9Tk7e5Tyv8Fv
Planens §3 sa at `ingest.materialize` er repoets ENE skrivepunkt på Door A, og
det premisset ble felt av måling FØR bygging: `materialize` er en ren delegasjon
til pinnet llm_ingestion_okf v0.3.2s `materialize_bundle`, som stager i minnet og
utfører sin egen disk-fase. Det finnes ingen callback mellom de to, så en gate
plassert der kunne bare kjørt ETTER at bytene landet — en opprydding, ikke en gate.
Sømmen ble i stedet kopier bundelen → materialiser inn i kopien → skann det som
ble generert → publiser eller forkast. Kopien er bærende, ikke bekvemmelighet:
bibliotekets §3 eierskaps-skann, kollisjonsgaten mot kuratert innhold og §6
index-merge leser alle den EKSISTERENDE bundelen. Staging i tom katalog mister
alle tre og publiserer en bundle uten kuraterte naboer — datatap forkledd som
sikkerhetsfiks.
De fire §4-beslutningene, tatt og målt: (1) ingen av guardens to preset —
Origin.EXTERNAL/AUTOMATIC, fordi trust_for utleder policy fra origin alene og
PRESET_USER_UPLOAD bærer en quarantine-semantikk Door A ikke har; (2) utfall per
BUNDLE, diagnostikk per DOKUMENT — delvis publisering ville etterlatt bundle +
index som svarer til intet manifest, men import_bundle itererer forbi første
avvisning; (3) Report til log.md, aldri konsept-frontmatter, der fire golden-suiter
pinner bytene; (4) mypy-override OG adapter, siden override alene gjør sømmen
type-blind i stedet for type-sikker.
`materialize` forblir ugatet med vilje — goldenene pinner den, og en kaller som
vil ha gaten ber om den ved navn.
Fem mutasjoner alle røde + grønn kontroll (hele suiten, ~120 s hver): detach
gaten · la den fyre ETTER publisering · Origin.INTERNAL · tom staging-katalog ·
rapporter kun første avvisning.
Målingen felte en VAKUØS test først: en hard injeksjon scorer fail_secure under
BEGGE trust-tierene, så Origin.INTERNAL-mutasjonen lot alle tre avvisningstestene
stå grønne — beslutning 1 så dekket ut uten å være testet. Båndet der tieren
faktisk avgjør er høy-entropi-innhold (quarantine_review vs warn), og testen ble
skrevet mot nøyaktig det før mutasjonen ble re-målt. Mutasjon 4 ble på sin side
felt av KUN én test; 809 andre merket ikke at bundle-kopien forsvant.
Laveste disposition er `warn`, ikke `allow` — `allow` finnes ikke i guarden. En
gate skrevet mot == allow ville avvist hvert dokument som noensinne ingestes.
Kriterium 5 står: demo-stdout er byte-identisk med tests/golden/demo-transcript.stdout,
målt både i suiten og ved eksplisitt kjøring. shared/ er urørt.
801 -> 810 passed / 4 skipped.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DDu94KoyxAmhJsG2n63X8Q
Kriterium 6 er selv-identitet: to kjøringer av en regredert demo er like enige som
to av en riktig. Fasiten forlater derfor prosessen. stdout pinnes ORDRETT (og er
dermed demoens abortsti); stderr normaliseres på nøyaktig to MÅLTE miljø-spann —
site-packages-prefikset og temp-katalogen — med po-sim- holdt synlig, fordi det er
en egenskap ved programmet og ikke ved miljøet. Pinnet stderr = fire linjer.
Kontrollen som forbyr at masken vokser er load-bearing: en droppende normaliserer
med fasiten regenerert under seg holder BEGGE likhets-testene grønne.
Pkt. 4: planens forhåndsskrevne frø-setning sa «én av de TO tidligere dommene».
Målt mot levert VEGLYS-bundle henter Kjøring B TRE — én fulgte med kunnskapsbasen,
to er demoens egne, én per tidsskala. Splitten avledes derfor fra kjøringen; en
håndskrevet «én av tre» ville vært den andre kopien som drifter.
Fem mutasjoner alle røde + grønn kontroll (hele suiten hver gang): ett byte i en
stdout-linje · detach dempingen · over-normaliser stderr · literal splitt · detach
frø-setningens print. Byte- og detach-mutasjonene ble fanget av KUN golden-testen;
den literale splitten av KUN skille-testen.
793 -> 801 passed / 4 skipped.
P3 a-c lukket. Pullen hentet commons 002f000+27cdce9: kriterium 8 grønt (nav-goldens
byte-uendret, målt både med git diff og shasum) og suiten uendret 785 — ingen abort.
Retningen på tallene er SNUDD (P3 b): manus-registerets kostlinjer er skrevet FRA
shared/examples/veglys-fv-soer/cost-baseline.json, ikke ved siden av den, og
baseline_from_scripted_candidate brukes ikke på denne stien — main() leser levert fil.
Overdrivelsen 2100000 er valgt utenfor commons' tall-inventar: målt fraværende fra
bundelen og over målt P90 1769915 (deres anslag var ~1770000). 600000/900000 ville
klarert gaten og aldri utløst Steg 5.
Målingen felte en defekt reserven skjulte: :g slår over i eksponentform ved 7. signifikante
siffer, så levert baseline printet 4.38615e+06. Reservens 300000 har seks siffer og nådde
aldri overgangen. _num erstatter :g begge steder.
Load-bearing MÅLT mot hele suiten, fem mutasjoner alle røde + grønn kontroll:
detach main-wiringen · reverter _num til :g · drift registeret ETT siffer (4386151 —
innenfor 5 %-toleransen, fanget av ingenting i 792 tester bortsett fra den nye) ·
sett flip_key til et token som finnes i bundelen · detach forankringen på bundle-stien.
785 -> 793 passed / 4 skipped.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BUjfw4eJdwwnqHSXhfcY6i
Two commands are now part of the install surface a fresh clone gets from `uv sync`:
`portfolio-optimiser` (run:main) and `portfolio-optimiser-demo` (simulation:main).
Deliberately two of five main()s — costsim/hitl/preflight stay module-invoked; every
name here is a name the freeze has to carry. Pinned against the INSTALLED distribution's
metadata, not the TOML: a [project.scripts] line that has never been synced is a claim,
not a command. Measured: stdout is byte-identical across both invocation forms.
stderr (P4 pkt. 2), the session's open decision, resolved by measurement rather than by
preference. Damped: the round-cap notice only, via a filter on the emitting logger, keyed
on the message and installed by main() — never at import, so a library consumer keeps its
own logging config. NOT damped: the two ExperimentalWarnings. They fire while the package
__init__ imports run -> agent_framework, always before simulation's own imports and under
both invocation forms, so silencing them would mean filtering warnings inside the library
package on every consumer's behalf; they are pinned in pkt. 3 instead. A console-script
wrapper was rejected for a second reason: the two forms would then write different stderr,
and a byte-fasit would pin the command rather than the program. stderr 6 -> 4 lines.
A first implementation wrapped simulation's own agent_framework import in a scoped mute.
Measurement showed it can never fire — the package __init__ has already imported
agent_framework by then — so it was removed rather than left as a green-but-dead seam.
Load-bearing MEASURED against the whole suite, five mutations all red + green control:
remove [project.scripts] · typo the target · detach the main() call · make the filter drop
everything · install the filter at import time. The typo mutation also felled a test: the
resolve-assert re-checked the expected constant against itself, and now resolves what the
distribution actually installs. 775 -> 785 passed / 4 skipped.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01C2bxLcCRguxXzpM4priTMn
The validator can reconcile a proposal against the project's actual cost lines
(S4.0 stage 0), but only when the knowledge base ships a cost-baseline.json —
and no bundle under shared/ has one. So on stage the gate reasoned only about
numbers the proposal supplied itself.
The reserve can never receive the file in shared/ (pull-only subtree, and demo
criterion 8 requires the goldens byte-unchanged). That is a placement
constraint, not an impossibility: materialize_anchored_bundle copies the bundle
and adds the file outside shared/, and the run path reads it through exactly
the seam a delivered bundle would use.
The baseline is DERIVED IN CODE from the scripted register, never typed beside
it — two sources of the same numbers drift, and drift is precisely what the
10 % probe models. On GO day the direction reverses (plan P3 b). Both scripted
replies must state the same cost lines or ValueError: were they to differ,
hypothesis #1 would be falsified by stage 0 instead of by P90 — the same
REJECTED line on screen, a different mechanism behind it.
10 % probe, measured: baseline x 1.10 -> FORKASTET at stage 0, before the
solver; corrected -> FORESLÅTT. Criterion 6 re-measured (stdout byte-identical
across two runs); stderr unchanged at 6 lines. The ONLY diff against the
un-anchored demo is the new KUNNSKAPSBASE block — everything else is
byte-identical, which is the problem: an anchoring nobody can see is one nobody
can check. Hence it is printed, and hence `provenance` is a required argument.
769 -> 775 passed.
Five mutations red + green control. The measurement failed the TEST first:
"ingen kostbaseline erklært" CONTAINS "kostbaseline erklært", and
ENERGI-TOTAL-EL already appears in the Step-2 line, so both assertions survived
the detach mutation. The two branches now share no wording.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GD6Y2Y23NZZxPYtSRoCmst
The Step-7 trace line said "lang fil-løkke" while the verdict arrived as a
function argument (`verdict_input`) — the short, in-run capture. The long loop
was tested but never exercised by the thing on stage.
An expert now drops a real verdict FILE (`write_verdict`) into an inbox between
the runs, and Run B is given `verdict_dir=`, so `run_project` merges it into the
store before the Step-1 fold.
Not done as the plan point was worded, and the difference is load-bearing:
routing the PERSONA verdict through the inbox would have put ONE marker on two
paths — Step 7 (inbox) and Step 8 (promotion) both end in Run B's prompt, so
either could carry it alone and `test_simulation_loadbearing.py`'s promotion
assertion would have stayed green with promotion detached. A second verdict with
its own marker keeps both seams independently red-able; `simulate_learning_loop`
raises when the two markers are equal. The inbox sits beside the bundle copy,
never inside it, and the id is an explicit sentinel (a minted id would collide
with the promoted verdict's, and `VerdictStore.add` is first-write-wins).
766 -> 769 passed (773 collected). Criterion 6 re-measured: stdout byte-identical
across two runs; stderr unchanged at 6 lines. Mutations measured against the full
suite, four red + a green control: detach `verdict_dir=` · point Run B at an empty
folder while the file is still written · marker set to `realization_rate: 0.82`
(measured present in the verdict seed) · marker set to `energy performance gap`
(measured present in a navigated concept file) · benign rename of the inbox dir.
Honesty limit found while measuring: the last two mutations fell on the causality
assertion, not the Run A control — generation prompts carry the debate output, not
the bundle context. The pair holds, but each assert defends a different property.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FVYDeJ9evZicgU5r3roZVW
The scripted proposer answered one hard-coded pair of proposals. A second project meant a second
hand-written selector, written under demo-week time pressure -- the risk the week plan names
explicitly (§4, risk 2). It is now a registry: `ScriptedCandidate` entries selected by
`scripted_proposer`, plus `project_id` as an argument to `simulate_learning_loop`.
The open decision was WHAT identifies the candidate in the prompt blob; the plan flagged it as
unverified, so it was measured. Two prompt shapes reach the selector: the debate prompt carries the
whole bundle context, the generation prompt carries `Project: {id} - {name}` plus -- as its context
-- the debate output, which is the selector's own earlier reply. So the cost code and the measure
name are present in the generation prompt only because the script put them there; keying on them
would key the script on its own output. The project id is the one identifier both shapes carry and
the framework stamps.
Validation, never repair: no match, or more than one, raises `ScriptedCandidateError`. A default
reply would answer an unregistered project with another project's numbers, which on screen is
indistinguishable from a correct run; an ambiguous blob is a data problem that must surface at the
rehearsal rather than be decided by registry order.
Load-bearing MEASURED against the whole suite, five mutations all red plus a green control: detach
the project keying - one global flip key - fall back on an unknown project - first-match on an
ambiguous prompt - detach the `project_id` argument. The flip-key test was rewritten mid-measurement
because its first form asserted on the FIRST registry entry, where "the matched candidate's key" and
"candidates[0]'s key" coincide -- it could not separate the two implementations, and proved nothing.
766 passed / 4 skipped. Simulation still exits 0, still prints eight labelled steps, still
byte-identical across two runs.
[skip-docs] README is deliberately untouched: O4 defers the README rewrite to 14-15 August, after
the demo has produced the evidence for the level-2 claim. CLAUDE.md carries the invariant.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XoHJCKBTjFKcjsfEQyGbzh
generate_via_llm consumed each validator Rejection internally (`last`), fed it into the
next attempt's prompt, and dropped it. So Step 5 was real but unobservable: a caller could
see THAT a proposal validated, never that it validated on attempt 2 after the deterministic
validator falsified attempt 1. It was the one step of the eight with no output to show.
The seam is a typed return value -- GenerationResult(outcome, refinements) -- rather than an
out-parameter or a callback: a returned value cannot be silently lost by a caller that forgets
to pass a collector, and mypy forces every call site to acknowledge it.
refinements carries ONLY rejections that were actually fed back. When the attempt budget runs
out the final rejection IS outcome; counting it here would be double-counting, and the bounded
control test goes red on the collect-everything implementation that gets this wrong.
The loop's bound is untouched: max_attempts and meter.tick_round stand, and `last` still drives
the prompt alone, so prompt growth is unchanged. run.py accumulates across _evaluate calls, so
_evaluate_mandate is untouched; RunResult.refinements defaults (the coverage precedent) and is
concatenated across approaches rather than keyed per approach -- stated as an honesty limit.
The simulation now shows it: the scripted proposer overclaims 250000, which the validator
falsifies against P90 = 90000, and the corrected 30000 validates. Only the overclaim is
scripted -- the rejection is computed. scripted_factory takes a per-role reply selector so this
needs no second scripted client body.
README records the two accuracy changes only (Step 5 is now inspectable; the simulation trace
shows the correction). The level-2 publishing claim stays deferred until after the demo (O4).
Load-bearing MEASURED against the full suite with a control, four mutations all red:
detach the returned history (4 tests) - collect-everything (control only) - detach the run
wiring (2 tests) - revert the simulation's proposer to a constant (the demo-protection test).
Control: 759 passed / 4 skipped; ruff, format and mypy clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017CcWFcREUi6YPjEpN3ACDP
The egress declaration (Trekk B3) says what a run MAY contact. It cannot say what
it DID: after the run, nothing distinguished "the agents queried the price
register" from "the agents ignored it", and a proposal resting on an external
service should be traceable to it.
ToolCallRecorder(FunctionMiddleware) mirrors BudgetMiddleware(ChatMiddleware) one
layer down — that one observes the debate's chat calls, this one its tool calls.
It observes only: call_next is always awaited, so a trace can never alter the run
it traces. The record lands on ProvenanceStamp.external_calls, read AFTER the
debate so it is a record rather than an intention.
MEASURED, not assumed, before any of it was written: FunctionMiddleware fires for
a tool served over a REAL MCP stdio subprocess, and context.function.name carries
the BARE tool name with no server prefix. That measurement decided the design —
MAF cannot tell us which server a tool came from, so attribution comes from our own
config, and a name allowed by two servers is recorded UNATTRIBUTED (server="")
rather than credited to the first match. Naming a service that may never have been
contacted is the one place a guess must not go.
Only CONFIGURED tools are recorded. The middleware fires for every function the
agents invoke, including the in-process retrieve_cost_docs on the road path;
logging those would turn the record into a false egress claim. An empty list is a
positive statement — nothing outside this process was contacted — which is why it
is always serialized rather than omitted.
Honesty limit, written on ExternalCall itself: this is the call and its source. It
is NOT evidence that the service's answer reached the proposal, nor a verified
rendering of that answer.
One finding, and it is the reason for measuring rather than trusting green: the
road-path negative test was VACUOUS. Its scripted tool call named an argument the
tool does not declare (code vs query), MAF rejected the call before invocation, and
the test asserted an empty record against a run where no tool ran at all — green
under the exact mutation it existed to catch. It now spies on the recorder and
asserts the invocation genuinely reached it before asserting it was not recorded.
This is last session's lesson again: a scenario that cannot distinguish two
implementations proves nothing.
The tool-call double is registered in test_scripted_client_consolidation.py's
_DELEGATING_OVERRIDES — it cannot live in the reply_selector seam, which returns a
reply STRING, and a response that is not text is its whole subject.
Load-bearing MEASURED (tests/test_b4_mcp_call_trace_loadbearing.py) against the
whole 755-test suite, four mutations all red: detach the recorder from the debate
middleware · record every function invocation · attribute an ambiguous name to the
first server · stop reading the recorder into provenance. Control: a run with no
configured servers records nothing, so the empty record is a real answer and not
the only one the seam can produce.
Ran it, not just tested it: the real recorder against a real MCP server subprocess
returns ExternalCall(server='prisregister', tool='lookup_unit_price'), and a
scripted CLI run's outbox artefact carries the empty list.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VtRd8y1PDPGwkrRXFhubqr
A run commissioned to evaluate three approaches wrote ONE proposal artefact, so
only the approach it selected could ever receive a verdict. The other two were
evaluated, reported in the settlement, and then taught the learning loop nothing.
The defect class is a key collapse, and it had two halves — fixing either alone
leaves it intact:
* the WRITER wrote one pair per run, so the non-selected approaches never existed
on disk;
* the READER (hitl._read_outbox_proposals) joins proposal to outcome on the
run_id FIELD read from file CONTENT, never the filename. Three files sharing
one run_id collapse onto one dict key, last write wins — so widening only the
filename would have produced three artefacts and still one pending row. This is
the S3.2 collision class: two rows under one key silently become one.
Artefacts are now keyed {run_id}-{approach_id}-*.json AND carry approach_id in the
payload; the join key is (run_id, approach_id). Two properties make them genuinely
judgeable rather than merely present:
* verdict_id is minted per approach (verdicts.verdict_key, the S3.2 content hash)
— reusing the run's single id would let one delivered verdict clear all three
from the queue;
* provenance.validator_decision follows ITS OWN approach — the run's stamp would
report a rejected candidate as validated, and nothing downstream could correct it.
verdicts.verdict_key is public so a run can stamp the key a verdict WILL arrive
under without capturing a decision nobody has made; it delegates to _mint_id
rather than restating the hash (the (p) rule: one keying rule, one copy).
The per-approach set REPLACES the run-level pair rather than joining it — the
selected approach is already among them, and writing both would count it twice in
hitl pending. The selected one carries the run's final outcome, so the outbox can
never disagree with the RunResult; the others carry the validator's verdict, the
only falsifier that ran on them.
mandate.py is deliberately untouched: hanging a ValidatedProposal off a coverage
row would drag validator — and pulp — into a module kept to pydantic+stdlib for
D7 portability, so _evaluate_mandate returns the evaluated outcomes alongside.
Ran it, not just tested it: a real CLI run wrote six artefacts and hitl pending
listed three rows. It also showed the honest edge — three approaches that produce
an identical candidate share one content-hash key, so one verdict settles all
three. That is correct (they were one candidate), and it is now documented.
Load-bearing MEASURED (tests/test_a5_per_approach_artifacts_loadbearing.py) against
the whole 750-test suite, five mutations all red: detach the per-approach writer ·
drop approach_id from the join key · reuse the run's verdict id · reuse the run's
provenance stamp · widen the filename but not the payload. Control: on a full
detach exactly the 5 new tests fail and 745 pre-existing ones stay green — the
no-mandate path is inert, and writes neither the filename segment nor the field.
Docs: bestille-en-kjoring.md (what the commissioner gets) + ekspert-svar.md (what
the expert's queue looks like, and that "rejected" is the validator's verdict on
the numbers, never a professional judgement of the idea).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VtRd8y1PDPGwkrRXFhubqr
Krav 3, and the operator chose the run path explicitly: the external service must
be reachable WHILE the run works, not only when documents are ingested. Until now
the run path had one in-process tool against a local folder — and on the bundle
path the agents had no tools at all.
MAF already ships the client (MCPStdioTool / MCPStreamableHTTPTool, verified in
the pinned 1.9.0 with allowed_tools and request_timeout), so `mcp_tools.py` owns
only what MAF cannot decide for us: which servers a run may contact, which of
their tools it may call, how long it waits, and where the credential comes from.
This is a DIFFERENT seam from ingest_mcp.py on purpose — that one pulls source
documents before a run and speaks to null-argument tools. Same protocol, different
job.
Every refusal is a live hazard, not tidiness. An empty allowlist would let the far
end decide what the agents may call, so naming the tools is mandatory. A
non-positive timeout is an unbounded wait against a third party. An unknown field
is refused rather than ignored, which is also what keeps a literal secret from
being parked in the config — there is no field for one, only the NAME of an env
var. A named-but-unset credential refuses instead of calling anonymously, because
an anonymous call can succeed with the wrong scope.
Egress is declared, always. Every server and permitted tool is named in the run
announcement before the first call — including when no --mandate is given, which
was a real hole: the announcement only printed with a commission, so configuring
servers without one would have contacted third parties with nothing printed at
all. --live-dry-run still opens nothing, because the tools are entered after the
dry-run cut: the promise to stop before the first call now covers egress too.
Threaded through BOTH modes. A flag accepted in one mode and silently dropped in
the other is the defect class this CLI refuses by name.
Load-bearing MEASURED against the whole 744-test suite, four mutations all red:
build the tools but never hand them to the agents (2) · never enter the
AsyncExitStack, so they are constructed and useless (1) · never declare the egress
(2) · drop the allowlist on the built client (1).
Two live docs claimed MCP was unwired in the run path; both corrected rather than
left to rot.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ULCqjLF61rehj5cZmdUoR3
`docs/bestille-en-kjoring.md` is the commissioning half of the expert-facing pair
(`ekspert-svar.md` is the judging half): the mandate file field by field, how to
run it, and — separated deliberately — what a commission does NOT do. It directs
what is evaluated, never what is approved.
Registered in _LIVE_DOCS, so it cannot silently fall behind the code.
The example output in it is COPIED FROM A REAL RUN, not composed, and running
that run is what found the defect fixed here: three approaches against the same
cost line each validated at 30000 NOK, and the settlement printed
"Validated total: 90000 NOK". Commissioned approaches are ALTERNATIVES — they
usually attack the same line — so summing them reports money the project cannot
realise. A domain expert reading that total would reasonably believe the run
found 90k.
The settlement now reports how many approaches held and which one the run
carries: a selection, not an arithmetic claim. That also removes the last money
addition from this module, which is the right place for it not to be.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ULCqjLF61rehj5cZmdUoR3
Krav 2: a run must be clear about what it shall do and achieve. `--mandate`
makes both ends explicit. BEFORE the first (paid) model call the run prints what
it was commissioned to do — objective, every named approach, whether its own
proposals are allowed, the scope, the caps, and which external services it will
contact. AFTER the run it settles: one row per approach, validated with the
figure, rejected with the validator's reason, or not evaluated with why.
The announcement is placed with the other refusals and above the scripted
banner, for the reason the required-args guard was hoisted there: a refused run
must not first print a banner about work it never did. `--live-dry-run`
announces without contacting anything, so a commission can be inspected before
it costs money.
`settle` renders the goal verdict it is GIVEN and never decides it. `ledger.to_ore`
is the framework's one NOK->øre conversion and the goal comparison already runs on
quantised integers, but `mandate.py` cannot import it without dragging `verdicts`
— and therefore agent_framework — into a deliberately framework-neutral module,
while a private copy of a money conversion is exactly the (p) defect. So the
caller decides and this renders; a goal figure without a decided verdict makes no
claim at all.
The caps the announcement prints come from named constants shared with
`run_project`/`run_portfolio`'s defaults — a second copy could drift and make the
announcement describe a run that never happened.
Load-bearing MEASURED against the whole 717-test suite, four mutations all red:
detach the announcement (4) · detach the settlement print (2) · make the CLI
loader tolerant (2) · announce the commission but never hand it to the run (2).
The last one is the one that matters: without it, a run could print a commission
it had no intention of executing.
DEVIATION from the approved plan, stated rather than quietly dropped: --goals is
still refused outside portfolio mode. Accepting it in single-project mode would
have admitted a flag whose documented function (the goal-stop against the ledger)
still does nothing there — the same accepted-but-inert defect --embedder-config
was just fixed for. The mandate's success_criteria carries "what shall this run
achieve" in the expert's own words instead; the numeric target stays portfolio-level.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ULCqjLF61rehj5cZmdUoR3
Carrying an approach into the prompt only half-answers krav 1. The expert asked
for their approaches to be CONCRETELY EVALUATED, which means each must reach a
verdict and each verdict must be visible. `run_project(mandate=...)` evaluates
every commissioned approach in turn — the run's own proposal last, when allowed —
each under the SAME meter. No new loop: the caps already in force are the bound.
`RunResult.coverage` is the settlement, one row per approach: validated (with the
figure), rejected (with the validator's reason verbatim), or not_evaluated (with
why). `not_evaluated` is the row that earns the type its keep — an approach the
run never reached must be reported as unreached, because an omitted row is
indistinguishable from an approach nobody ordered. That silence is the defect
class krav 1 is asking us to remove.
Budget exhaustion mid-list is reported, not swallowed. But if the FIRST approach
exhausts it there is nothing honest to return, so BudgetExceeded propagates
exactly as before — a run that produced nothing must still fail loudly.
RunResult stays single-outcome (portfolio aggregation, outbox artefacts and HITL
keying all rest on that). The choice is deterministic: highest validated saving,
ties by mandate order — never whichever ran last.
Load-bearing MEASURED against the whole 702-test suite. FIVE mutations red:
evaluate only the first approach (5 red) · drop the rejected rows (3) · ignore
allow_own_proposals (1) · select produced[-1] (1) · select produced[0] (1).
The sixth measurement is why this commit exists in this shape: the ordering
mutation FIRST STAYED GREEN. The test had placed the bigger approach last, where
"highest saving" and "whichever ran last" give the same answer, so an
order-dependent implementation passed it. A scenario that cannot separate two
implementations proves nothing about either — the test now pins BOTH orderings,
and each mutation direction fails one of them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ULCqjLF61rehj5cZmdUoR3
Operator feedback: fagpersoner must be able to name the approaches a run shall
evaluate for a project, and/or ask the system for its own. Today the hypothesis
prompt is hardcoded ("Propose ONE concrete cost-saving measure") and the only
expert-facing lever, --dimension-config, FILTERS what may pass the scoping gate
rather than DIRECTING what is spent attempts on. This is the input that was
missing.
`mandate.py` is the typed commission + a fail-fast loader (mirrors
`load_dimension`/`load_goal_config`): missing or malformed refuses, because a run
must never proceed on a silently degraded commission — the coverage report would
then describe work nobody ordered. Stdlib + pydantic only, so it joins
`_MAF_FREE_MODULES` and can be mirrored to the D7 sibling.
Two refusals carry real defect classes: an EMPTY commission (no approaches and no
own proposals) is a caller error, not a result; and a duplicate approach id — or
one claiming the reserved OWN_PROPOSAL_ID — would collapse two coverage rows onto
one key (the S3.2 key-collision class), which is exactly the silence the coverage
report exists to prevent.
The numeric target is deliberately NOT duplicated here: it already lives in
GoalContract, and two copies of one number drift apart ((p) precedent). The
mandate carries intent; `announce` merely restates the figure.
`_build_messages(approach=...)` switches the opening instruction from *find one*
to *quantify THIS one*, carrying the expert's label and description VERBATIM —
the description is the reason the approach is worth trying, the one part the model
cannot infer from cost data. `approach=None` is byte-identical to the previous
prompt, so every existing run and golden is untouched.
The gate is unmoved: `validate_proposal` is called exactly as before. A
commissioned approach gets no discount — the expert directs what is EVALUATED,
never what is APPROVED.
Load-bearing MEASURED against the whole 695-test suite, four mutations all red:
detach the approach injection (2 red, control stayed green) · let a commissioned
approach bypass the validator · make the mandate loader tolerant · drop the
empty-commission refusal.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ULCqjLF61rehj5cZmdUoR3
CORRECTION TO b33ea00. That commit message called the v0.4.0 behaviour a
regression that "breaks our §6 removal path". That claim is WRONG, and this
commit is the record of it — the history is not rewritten.
What v0.4.0 actually introduced is §10.2 per-manifest ownership.
`_is_ingest_owned` (materialize.py:133-160 at v0.5.0a2) compares the
`ingest_manifest` reference's STEM against the running manifest's stem, and
its own comment says why the stem rather than the whole stamp: an EDITED
manifest (new sha -> new stamp) must still reclaim the files its previous run
wrote, while a DIFFERENT manifest sharing the bundle keeps its own.
Our fixture named manifests `manifest-{len(extractions)}.json`, so "the same
manifest, edited" silently became `manifest-2.json` -> `manifest-1.json` —
two stems, i.e. two rival manifests. Refusing to overwrite was CORRECT
behaviour. The fixture was written when ownership was manifest-agnostic and
the naming was pure convenience; v0.4.0 made that convenience load-bearing.
`_write_project` now takes an optional `filename`, and the re-ingest case
pins one stem across both runs, so it means "an edited manifest" on every
version rather than by accident.
MEASURED both ways: green at the pinned v0.3.2 (the distinction is invisible
there), and the WHOLE suite green at v0.5.0a2 — 668 passed. So there is no
technical blocker to the version move at all. What remains is not technical:
1. v0.4.0+ makes `llm-ingestion-guard>=0.2,<0.3` a hard runtime dependency
(v0.3.1/v0.3.2: `dependencies = []`), which moves two invariants this
repo publishes. Operator's call.
2. v0.5.0a2 is an alpha its own CHANGELOG scopes to a named pilot set this
repo is not in. llm-ingestion-okf's call; asked via coord.
The false report was corrected upstream the same hour it was sent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GyAbxJoyypnLLUDcMvnKh8
The pin had sat at v0.3.1 with STATE calling the hold "deliberate" and
recording no reason. Measured: no coord message ever announced v0.4.0 or
v0.5.0a* to this repo, so the hold was drift wearing a decision's clothes.
v0.3.2 is a pure fix (frontmatter and index labels emit verbatim; only
source_query is whitespace-collapsed, per ingest-spec §5), keeps
`dependencies = []`, and is green here: 668 passed.
WHY NOT FURTHER, both measured rather than assumed:
1. v0.4.0 introduces a REGRESSION that breaks our §6 removal path.
Bisected v0.3.2 OK / v0.4.0 RED with a minimal repro: materialize a
bundle, then re-materialize it with a CHANGED manifest, and the library
no longer recognises its own stamp —
MaterializationError: generated filename 'ingest-costs.md' collides
with an existing file that does not carry the ingest stamp
The stamp carries the manifest's name+hash (`ingest_manifest: m2@…`), so
editing a manifest makes every file it previously wrote look curated.
Re-ingesting the SAME manifest is fine, which is why fixtures miss it.
It is `tests/test_ingest_loadbearing.py::test_reingest_with_active_
removal_preserves_promoted_and_curated` that catches it. Reported
upstream; not ours to fix.
2. Everything past v0.3.1 adds `llm-ingestion-guard>=0.2,<0.3` as a HARD
runtime dependency (v0.3.1/v0.3.2: `dependencies = []`). That flips two
documented invariants here — pyproject's "zero runtime deps" comment and
the STATE marker line the guard repo reads machine-readably ("not a
runtime dependency today"). An operator decision, not a version bump.
3. v0.5.0a2 is an alpha whose own CHANGELOG scopes it to a named pilot set
— portfolio-optimiser-claude, the marketplace catalog, claude-code-llm-wiki
— and says "do not pin this tag outside the pilot set", with the v0.2
surface free to change without a deprecation cycle. This repo is not a
pilot. Joining is llm-ingestion-okf's call, requested via coord.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GyAbxJoyypnLLUDcMvnKh8
Walked from a fresh clone: `--embedder-config` was accepted in every mode
without `--semantic-retrieval` and then had no effect whatsoever. MEASURED,
not inferred — an injected embedder is consulted ZERO times with the flag
off and once with it on, because the only consumer is the HybridRanker that
flag builds; the default StructuralRetriever takes no embedder at all.
That is the silent-ignore this CLI's flag contract exists to prevent, and
the same ground on which `--semantic-retrieval` itself is already refused
when it cannot take effect.
REFUSED, not wired — the opposite call from `--scripted-replies` in
portfolio mode, and for a stated reason: there the seam already existed, so
refusing would have left a whole mode without an offline door. Here there
is nothing to wire to.
Mode-independent (both modes gate the embedder on the same flag) and placed
ABOVE the scripted door, mirroring the required-args hoist: a refused run
must not first print a banner claiming a scripted loop closed.
Five mutations against the WHOLE suite, all red, each isolating one seam:
detach the refusal (3 red) · scope it to single-project mode (portfolio arm
red) · move it below the banner (banner arm red, rc intact) · build the
ranker unconditionally (the zero-consultation measurement red) · ignore the
injected embedder (its control red).
663 -> 668 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GyAbxJoyypnLLUDcMvnKh8
An independent reviewer found it and the claim was verified by measurement
before being accepted, not taken on trust.
test_portfolio_scripted_pass_makes_no_real_client raised from the patched
_default_factory to prove the scripted portfolio pass never builds a production
client. It cannot: the factory is called inside run_project's coroutine, and
run_portfolio gathers with return_exceptions=True, so the AssertionError was
collected into a RunFailure and never escaped. Measured directly -- with BOTH
the client_factory wiring and the rc rule detached, the blade stayed GREEN. It
was going red on rc alone, which means it was testing defect B while claiming
to test defect A. The original seven-mutation sweep did not catch this because
each mutation was applied singly, and dropping the wiring alone still flips rc.
Replaced with a call sentinel: a list appended inside the factory and asserted
in the test body, which the wave handler cannot swallow. Re-measured -- red on
the wiring detach alone, and red on both detaches together.
The docstring now also states what the sweep could not: blades 1 and 8 are
environment-conditional. The local profile points at loopback, so on a machine
running a local model server the wiring detach would make real calls and could
complete the pass. Their red was real on the machine it was measured on and is
not portable; blade 2's is.
Separately, the budget-stop print is marked as defensive and currently
unreachable from main(), because main() never constructs a PortfolioMeter and
every write to budget_stop is gated on one -- the strict=True precedent
directly above says untested future-proofing must be labelled as such. The
README claim that a portfolio pass reports a cap stop is corrected to say the
cap has no CLI flag yet. Noted for whoever wires that door: BudgetRefused is a
RuntimeError and the existing except clause would not catch it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0118noV9rCfrdREH26XqZB5z
The operator is not a domain expert, so the domain content is mine to own --
and the one thing the loop asks a human for is exactly the thing no example
existed for. docs/ekspert-svar.md is written for whoever has to deliver the
verdict: the two forms a judgement can take (a --rationale string during the
run, a JSON file in the inbox for later runs), where each field comes from, and
four complete paste-ready answers.
Every command and every verdict in it was RUN from a fresh clone before it was
written. The `hitl pending` line quoted is verbatim output. The rejection
answers close the gap STATE has carried since the demo shipped: the README
shows the VALIDATOR refusing a number, but nothing showed an EXPERT refusing a
proposal whose numbers are fine -- the only judgement in the whole loop that a
machine cannot make. Two rejection shapes are given, because "not feasible
here" and "right measure, wrong cost base" teach the system different things.
Everything is marked AI-authored and not verified professional judgement.
Also corrects the --outbox-dir help text, which claimed sharing a folder with
--verdict-dir "re-ingests raw agent output past the Step-8 promotion gate".
Measured, by pointing both at one folder and running twice: it does not. The
outbox artefacts are named {run_id}-*.json and carry none of the verdict keys,
so the tolerant inbox loader skips them and the run is unaffected. The hazard is
real but latent -- a future verdict-shaped artefact in the outbox -- so the
warning stays and says what is actually true. This also answers STATE's open
question about enforcing the distinction in the CLI: no. There is no reachable
contamination to refuse, and a guard for an unreachable case is the kind of
error handling this repo declines to write.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0118noV9rCfrdREH26XqZB5z
Walking --portfolio end-to-end as a downloader would -- which no session had
done -- surfaced two defects of a class this repo already legislates against.
(A) --scripted-replies was silently DROPPED in portfolio mode. main()'s
portfolio dispatch returned before the block that builds the scripted client
factory, and the flag was absent from the single_only refusal set: neither
honoured nor refused. Measured against the shipped reference portfolio: no
banner, four real model calls attempted, four APIConnectionError. The previous
session joined this flag to the --report allowlist and missed the portfolio
partition. Resolved by WIRING rather than refusing -- run_portfolio already
exposes the same client_factory seam, and refusing would have left portfolio
mode with no offline door at all for an adopter without a model budget. The
scripted block is hoisted above the dispatch; the single-project required-arg
and semantic-retrieval refusals are hoisted with it so an incomplete argv is
still refused BEFORE the honesty banner could claim a scripted run happened,
and the refusal order within single-project mode is unchanged.
(B) A portfolio pass reported one of its four outcome channels. failures
(S3.3 collect-and-continue) and budget_stop (S3.4 global cap) never reached the
operator and rc was unconditionally 0, so the four-failure pass above printed
NOTHING and exited 0 -- silence read as success. BudgetStop is a separate field
precisely so exhaustion can be told from success; the CLI showed neither.
Failures now print to stderr with project id, error type and message; the
budget stop prints its four numbers; rc is 1 iff something raised. A budget
stop alone stays rc 0: exhaustion is a structured stop the operator asked for
by setting a cap, not a crash. Completed runs still print, so the non-zero rc
does not undo collect-and-continue.
Load-bearing MEASURED against the whole 662-test suite, seven mutations all
red, including both controls: detach the client_factory wiring - make the
banner a single-project courtesy again - detach the failure print - revert rc
to 0 - detach the budget-stop print - print the failure line unconditionally
(control) - print the budget-stop line unconditionally (control).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0118noV9rCfrdREH26XqZB5z
Operator ruling 2026-08-05, which settles decision (g): planning documents are
generally never public, and what OUR OWN sessions generate does not go out on
the forge at all. The example itself stays public so others can run the
process.
`.claude/projects/` is the Voyage session workbench -- 25 briefs/plans/reviews
this project's own sessions produced. Untracked and gitignored, exactly as
STATE.md already is, and for the same stated reason: this repo has a public
mirror, so that class of material is local-only rather than tracked.
The line is drawn at who wrote the document, and it is drawn deliberately:
`docs/plan/`, `docs/research/` and `docs/rapport/` stay tracked. Those are
curated, dated documents written for the repo's readers, three of them linked
from the README as the decision record. Move that line if it was meant wider.
Two files were NOT process artifacts and are not deleted. Both
`build_fixture.py` scripts are cited by tracked tests
(`test_ingest_golden_sql.py`, `test_ingest_golden_http.py`) as the documented
rebuild path for byte-exact goldens -- reproduction code that had landed in the
wrong directory. Moved next to the goldens they build; both docstrings updated,
so no tracked file is left pointing into an untracked tree (verified: the only
remaining `.claude/projects` string in a tracked file is the .gitignore rule
itself). One prose reference in the dated Foundry auth recipe was dropped for
the same reason.
652 tests still pass.
Does NOT address the 27 of these already readable on open/ since the S12
release -- untracking stops future publication only. That retraction is a
separate operator decision and is deliberately not taken here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWsexbQjPo9rsV3aUE54ZS
An adopter without an API budget had two half-doors and no whole one.
`--live-dry-run` takes their own bundle but stops before the first model call
(`run_project` returns a DryRunReport), while `portfolio_optimiser.simulation`
runs the complete loop but only over ITS bundle with ITS scripted answers.
The seam for the missing third case -- the whole loop over your OWN data,
offline -- already existed as `run_project(client_factory=...)` and had zero
CLI exposure. This is the door onto that one seam, not a second implementation
of it (`scripted_factory` is imported lazily; `simulation` imports `run`, so a
module-level import would be circular).
The honesty banner is part of the feature, not decoration (maalbilde §1): a
scripted run that reads like a model run is worse than having no offline mode,
so every scripted invocation prints what is real (context navigation, debate
plumbing, deterministic validator, verdict) and what is not (the answers).
The two offline modes are mutually exclusive rather than one silently winning,
`--report` mode refuses the new flag by allowlist, and a replies file that
cannot serve the run is refused at the door rather than surfacing as a KeyError
mid-run.
Load-bearing MEASURED against the whole suite (645 -> 652), six mutations all
red: detach the wiring · detach the banner · detach the dry-run exclusivity ·
drop the flag from the --report allowlist · make the loader tolerant · control
(print the banner unconditionally).
The --report blade was measured GREEN first: with a non-existent ledger path
the load failure refused before the gate and masked it entirely. Rewritten
against a valid saved ledger, so rc 1 can only come from mode-exclusivity.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWsexbQjPo9rsV3aUE54ZS
Two rules existed. verdicts._unquote took quotes off correctly; the
bundle_context title renderer stripped only `"`. Measured before the fix:
`title: " Spaced "` rendered as `## concept: Spaced ` (the whitespace
half kø-(a) named), and `title: 'Single'` rendered its quotes verbatim —
both ordinary YAML a hand-authoring curator writes, and both reach the
agent's read-context.
The (p) defect class: a duplicated conversion drifts, and the drifted copy
decides something. Same fix shape — ONE source, owned by the module that
owns parse_frontmatter. okf.unquote_scalar is now the rule; verdicts
delegates by identity, so the structural key that _mint_id hashes is
unchanged.
The commons-owned nav-goldens could never have caught this: every golden
title is double-quoted with no inner whitespace, so both rules render them
byte-identically. That is asserted as a control, and it goes RED if a
future golden gains a discriminating title.
Load-bearing MEASURED against the whole suite, three mutations:
- weaken the renderer back to .strip('"') -> ONLY the 2 new rows red,
643 others green (incl. nav-goldens) = it covers ground nothing did
- reintroduce a private _unquote copy in verdicts -> only the identity
test red, 644 green (the copy is behaviourally identical TODAY, which
is exactly why identity is the only thing that catches the class)
- (control, kø-(n)) add `import numpy` to a second src module -> the
existing test_semretrieval_is_the_sole_numpy_importer goes red alone,
confirming that gate is live rather than green-but-dead
638 -> 645 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DVMsih6tXx39VMyq6wJ5H7
docs/extending.md claimed SEMANTIC_WEIGHT_DEFAULT = 0.5 long after the code
lowered it to 0.25. It was found by accident while editing the neighbouring
line; nothing in the suite would ever have caught it. This is that gate.
Both bounds were MEASURED, not assumed:
- SCREAMING_CASE discriminates exactly. Across every numeric `name = value`
code span in docs/ it selects the 2 real module constants and rejects all
10 kwargs/locals (max_attempts=3, concurrency=3, realiseringsgrad=0.79).
Ordinary prose stays freely editable, so the gate has no reason to be
switched off.
- Dated documents are observations, not contract. A spike finding or a July
review records what was true when measured; rewriting it to track the code
would falsify the record.
Measuring also rewrote the ambiguity rule: SEMANTIC_WEIGHT_DEFAULT is bound
in both semretrieval and run (a re-export), so a "same name in two modules"
check would have been RED on today's code. Only DIVERGENT values are refused.
Fail-closed throughout, per write_concept_file / read_spend: an unknown
constant name is an error rather than a skip, and a document that is neither
listed live nor recognisably archived goes RED asking to be classified —
otherwise a new guide would be silently unguarded.
Load-bearing MEASURED against the whole 638-test suite:
- drift the DOC (the original defect) -> only this gate goes red; the other
637 stay green, so it covers ground nothing else did
- drift the CODE -> this gate and the semretrieval weight gate both go red
- make an unknown name tolerant -> red
- drop the only citing doc from the live list -> red (twice: coverage and
classification)
- add a new unclassified guide -> red
- (control) remove this gate entirely, code still drifted -> the adjacent
semretrieval gate still goes red, so nothing is masked in either direction
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01D9AAyWtMqr4HjftKaegTtS
`cosine`'s docstring claimed its guard was load-bearing because "a NaN reaching the
ranking sort key would corrupt ordering silently rather than failing loudly" — but the
guard tested `norm == 0.0` only, which a NaN or inf norm passes straight through. The
claim was prose, not behaviour.
Measured, not assumed: `cosine(unit, nan_vector)` AND `cosine(unit, inf_vector)` both
returned `nan`, and a NaN sort key made ranking INPUT-ORDER-DEPENDENT — six permutations
of the same three candidates produced four distinct orderings. That defeats the total
order `HybridRanker` documents ("`id` makes the result independent of input order").
Refuse rather than coerce, and deliberately NOT symmetric with the zero-norm branch: a
zero vector is a legitimate handled state (`FakeEmbedder` returns `np.zeros` by design),
whereas a non-finite component only ever means the INJECTED embedder is broken. Scoring
it `0.0` would launder that into "no semantic similarity" while ranking proceeded on a
forged signal — validation, never repair, mirroring `read_spend`.
Reachable via the documented `Embedder` extension point, not the shipped fake; scoped to
the norms (90% principle — a finite-normed dot-product overflow is not chased).
Also corrects `docs/extending.md`, which stated `SEMANTIC_WEIGHT_DEFAULT = 0.5` while the
code has said `0.25` since the weight was lowered.
625 -> 630 tests. Load-bearing MEASURED against the WHOLE suite, five mutations all red:
detach the guard entirely · coerce to 0.0 instead of raising · check only the first norm ·
drop "non-finite" from the message · (control) detach the zero-norm branch, which fails
ONLY the zero-norm test — the new guard does not mask the existing one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018V9vNBmxAmgJ2JMoHByiHS
Advisor review of the prior commit (5269b7d) found the except TimeoutError
branch was wider than the brief asked for: builtin TimeoutError is also
socket.timeout (3.10+) and asyncio.TimeoutError (3.11+), so any TimeoutError
reaching that clause got relabeled mcp_timeout regardless of source. Gate on
anyio.CancelScope.cancelled_caught instead, mirroring _unwrap_ingest_error's
ownership rule (own it, wrap it; otherwise, untouched).
Measured: no live trigger exists today (MCP's own internal read-timeout
converts to McpError before reaching us; a server-side TimeoutError becomes
an ordinary isError result) -- pinned with a synthetic test raising from
StdioServerParameters construction, inside our fail_after scope but before
either nested task group, so it arrives ungrouped. Four mutations red against
the full 625-test suite: drop the translation, revert to asyncio.wait_for,
relabel the code, and drop the cancelled_caught gate.
Also promotes anyio to a declared direct dependency (was transitive via mcp
only) -- ingest_mcp.py now imports it directly.
asyncio.wait_for cancelling stdio_call_tool's run() from outside the anyio
task groups it awaits (stdio_client, ClientSession) never surfaced a
TimeoutError: measured against a real hanging server, the mismatch produced
an anyio.BrokenResourceError wrapped in a BaseExceptionGroup instead. Moving
the deadline to anyio.fail_after, nested inside both task groups, lets
anyio tear down its own structure cleanly and raise a plain TimeoutError,
which is now translated into IngestError(code="mcp_timeout") alongside the
mcp_tool_error/mcp_non_text_content family.
BudgetExceeded carries kind/limit/observed as ONE structured stop event, but only
`observed` was undefended. Measured against the whole suite before writing anything:
four of five raise sites (TokenMeter.charge, tick_round, and BOTH arms of exhausted())
could report any value at all without a single one of 621 tests noticing. Only
PortfolioMeter.check was covered.
What hid it: spikes/_harness.py carries its OWN copy of BudgetExceeded/TokenMeter, so
the spike suite's `observed` assert never touched the shipped module — the production
tick_round had no direct test whatsoever.
exhausted() is the only site that CHOOSES a ledger (the S3.4 pre-call guard), so a
refusal naming portfolio_tokens while reporting the run's own spend would misdirect
every reader of it. Both arms are pinned with observed != limit on purpose: at
exactly-exhausted the two coincide, and a test written there would pass on an
implementation that echoed the cap back as the spend.
No defect in the values themselves (unlike kø-x and kø-p) — the triple was coherent at
all five sites; the gap was purely coverage.
Load-bearing MEASURED: nine mutations, all red — five observed mutations (including the
control) and four echo mutations. 621 -> 623 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LW749xcXQmVEgdipB6KNm4
Two quantization orders existed and met at exactly one comparison.
SavingsLedger quantizes every realized candidate to integer øre and sums the
ints; run.py's goal baselines summed Project.total_cost FLOATS across items and
projects and quantized the total once. _goal_limit_if_reached compared the
former against a threshold derived from the latter — so whether a portfolio pass
stops early was decided by two differently-computed sides.
Measured divergence: three 60000.005 NOK lines are 18000003 øre quantized first
but 18000001 summed first (the float sum drifts to 180000.01499999998).
Decision: quantize per cost line, then sum integers. Each CostItem IS a money
amount — S4.0 made per-line quantity/unit_cost the validator's ground truth — and
integer addition is associative, keeping totals order-independent under the D-D
wave model, which the float fold is not.
ledger.to_ore is now the framework's one NOK->øre conversion; run.py imports it
rather than keeping a private copy (the S4.0 REPLIES precedent).
Measuring the mutations found two further gaps, both now closed: the per-project
baseline is a SECOND call site whose mutation survived the whole suite, and
realize bypassing to_ore with a raw float*100 was caught by nothing.
Load-bearing MEASURED (tests/test_money_quantization_loadbearing.py), five
mutations all red: detach the portfolio baseline · detach the per-project
baseline · reintroduce a private copy in run.py · change the rounding mode · let
realize bypass to_ore. 615 -> 621 tests.
Honesty boundary: sum_claimed_saving_nok (run.py:_aggregate) is deliberately
untouched — a float NOK reporting field that is never quantized and never
compared against the ledger, hence outside the ordering defect.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WiY53sm8JFqk7NN75g5wRS
`stdio_call_tool` shipped never having been executed end to end — docs said so
explicitly. Running it found a real defect: `stdio_client` and `ClientSession` are
each an anyio task group, and anyio re-packages anything leaving one in a
`BaseExceptionGroup`. Both errors the transport raises from inside the session
(`mcp_tool_error`, `mcp_non_text_content`) therefore reached callers as exception
groups, never as the `IngestError` the whole Door A path catches and switches on by
`code`. No canned-tool test could see this: they never enter a task group.
`_unwrap_ingest_error` recovers the owned error and re-raises it; anything unowned is
re-raised untouched, so this narrows an exception group rather than blanket-catching.
Duck-typed on `.exceptions` because `except*`/`ExceptionGroup` are 3.11+ and this
project supports >=3.10.
Verified against a REAL server subprocess (a local process costs no model tokens, so
the repo's cost discipline is untouched; the contract tests still spawn nothing):
`examples/ingest-golden-mcp/` + `tests/test_ingest_golden_mcp.py` — byte-identical
golden extraction mirroring the http/sql goldens, plus the tool-error and
missing-`server_ref` branches.
Also recorded: a server on the ingest path must expose a NULL-ARGUMENT tool, so
`datasource.build_mcp_server` cannot serve it (`retrieve_cost_docs(query)` has a
required parameter, verified to return an error result). The two are separate seams
by design.
Load-bearing MEASURED, five mutations all RED: detach the unwrap · detach
`initialize()` · make the error code generic · detach the `isError` branch · change
one byte of the served body.
612 -> 615 tests. ruff + format + mypy clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WiY53sm8JFqk7NN75g5wRS
Every stage of validate_proposal reasoned only about numbers the proposal itself
supplied, so an internally-consistent hallucination cleared the whole gate (F3).
A new stage 0 reconciles each affected_item against the project's CostBaseline
before the CBC solve: an unknown cost code is rejected, and a real code carrying
a quantity/unit_cost outside the configured tolerance (5% default, relative to
the baseline value) is rejected. Validation, never repair.
The baseline argument is OPTIONAL (None = pre-S4.0 behaviour), but both run
paths set it: the road path projects project.cost_items, the bundle path loads
cost-baseline.json when the bundle ships one. Bundles written before the
amendment stay un-anchored, so the commons-owned goldens run byte-identically;
a baseline that exists but is malformed still raises on both loaders.
F8: the method-specific cap now comes from the METHOD_CAPS registry (measure
type -> fraction, injectable) instead of an energy_efficiency string comparison.
The baseline format and tolerance semantics were decided locally — the commons
amendment (D-A pt. 2) never arrived, exactly as in S3.2. D7 mirroring stays open.
Three portfolio fixtures quoted cost codes belonging to OTHER projects; the new
gate caught them. They now quote each project's own lines, and the two copied
REPLIES tables import the single source instead of drifting from it.
Load-bearing measured (tests/test_s40_cost_baseline_loadbearing.py), six
mutations all red: detach the reconciliation stage; detach the magnitude
tolerance; detach the road wiring; detach the bundle wiring; ignore the injected
cap registry; make the optional loader tolerant of malformed content. Control:
with the road wiring detached the repaired portfolio fixtures still pass, so
they are not masking the seam. 597 -> 612 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JdwK7bQ4BZkWH4t8MRDKb4
seed_store_from_bundle keyed EVERY `type: verdict` file on bundle_candidate_features — the single
candidate the bundle's validator-input.json describes. A bundle carrying verdicts about several
candidates collapsed them onto one key, so a verdict about candidate B scored a perfect structural
match against candidate A's query and could be folded into A's hypothesis prompt. The ExpeL
substrate was single-candidate by construction.
A verdict file may now carry its own structural key in frontmatter (affected_codes / measure_type /
claimed_saving_nok); absent, keying falls back to the bundle candidate, so every pre-S3.2 seed keeps
working unchanged. promote_verdict writes the three fields, so a promoted verdict — frequently about
a different candidate than the target bundle's projection — does not impersonate that candidate.
Semantics decided HERE, not pulled: commons' seeding rule (method-spec §3 Steg 1 + bundle example)
has not arrived; we said we would build locally first. D7 mirroring stays open.
- ALL THREE fields or none. A partial declaration raises VerdictFrontmatterError rather than merging
with the bundle candidate, which would mint a key belonging to NEITHER candidate. Validation,
never repair (mirrors write_concept_file); the tolerant-skip rule belongs to the RAW inbox layer.
- claimed_saving_nok parses via json.loads — the SAME literal rule the IR projection went through —
and is written back with str() of the raw value. _mint_id hashes that value, so 30000 and 30000.0
are different keys; a normalising writer would split one candidate's signal across two ids.
- The structural key is signal-free, so it does not weaken the Step-8 no-leak property (Test C green).
Load-bearing MEASURED, five mutations all red: detach per-verdict keying · detach the fields
promote_verdict writes · make a partial/unparseable key tolerant · normalise the magnitude on write ·
remove the fallback (control — breaks the step1 suite at collection, proving the fallback bears load).
589 -> 597 tests. Full gate green (pytest, ruff, mypy).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QkjvTTxrg9LTrmghebfiij
Two tightenings, each measured by a detached-mutation run:
(1) The validator now blocks a claim above the CBC nominal feasible, in ADDITION
to the P90 stage. Neither dominates the other: an upward-skewed assumption band
lifts P90 ABOVE nominal -- so P90 alone passed review counterexample #1 (claim
100k, nominal 90k, band [0.70, 1.40], measured P90 121057) -- while a
downward-skewed band pushes P90 below it. Independent gate, same Rejection type,
existing rejections keep their existing reason.
(2) An assumption band must enclose its item's unit_cost (low <= unit_cost <=
high, inclusive). A band that misses it states a different price rather than an
uncertainty, and every Monte Carlo draw would then sample away from the item's
stated cost. Checked exactly where the Monte Carlo looks bands up -- per affected
item, by code; a band keyed to no affected item is never sampled and so has no
unit_cost to enclose.
The premise was re-verified against ground truth before building on it, not
taken from STATE: 05.2 unit_cost 215 in (200,230), 03.1 310 in (290,330),
ENERGI-TOTAL-EL 1.0 in [0.70,1.40] and (0.8,1.2). No fixture violates it.
The LLM path already catches ValidationError as a meter-bounded retry
(generate.py:138), so the new invariant cannot crash a run.
Mutations, all RED: detach the nominal block; drop the model_validator
decorator; make the enclosure strict. tests/test_bygg_energi_mikro.py and the
commons golden are UNCHANGED and green -- the regression proof.
586 -> 589 tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017DDXwUqHVAQQeYE7X1TXy5
