open/portfolio-optimiser
1
0
Fork 0
Code Issues Releases 3 Activity
portfolio-optimiser/docs
History
Kjell Tore Guttormsen db38cfc1a4
 feat(round-builder): one command turns a run's outbox into a round the gate can read [skip-docs] 
python -m portfolio_optimiser.evals.round_builder --outbox <dir> --round <n> --ran-at <ISO>
writes <rounds-dir>/<n>/ with the run's artefacts COPIED in, outcome.json derived from that
copy, and report.md -- the one artefact in a round a domain expert reads and corrects. Round 0
of the v1 criterion can now be made; it counted 0 of 3 because it could not be, which is a
different failure from a round nobody had held.

What it derives it derives with the gate's own functions rather than a second copy: verify_run
decides whether the run stands up to itself (an artefact contradicting its coverage row, a
half-missing family and a stray artefact are all refused AT THE SOURCE, before a byte is
written), stage_of gives column (c), row_changed gives the report's "changed since the previous
round", parse_time refuses a stamp without a zone, safe_rounds_dir refuses a round directory the
repo would commit. The validated total is ledger.to_ore per amount, summed as integers.

Two things it never does, and both are the point. It never writes the operator's attestation --
the gate stops at FORM OK without one, and that is correct, because no arrangement of files can
witness that a run happened. And it never invents: --ran-at is required because no outbox
artefact carries a clock, and feedback_ids stays empty because no run records which feedback
item produced which row. The report says "ingen tilbakemelding forklarer dette" on every changed
row rather than hiding that model noise and an answered objection look alike.

Chosen and why: --ran-at as a required argument rather than the coverage file's mtime, because
an mtime is a filesystem attribute one call sets and reading it as evidence made row 2 green on
a tree nothing had run in (18.09). The report carries no raw stage identifier -- every stage
sentence is "<short name>: <explanation>" so the one-line diff of what changed has words a
reader can act on. A citation shows its COUNT, because a run that cited 446 places and one that
cited one must not look the same.

[skip-docs]: the ledger row is in docs/invarianter.md, which is where this repo's rules live.
README is the product's front door and this is an operator tool behind `python -m`, the same
class as costsim/hitl/preflight, which README deliberately does not carry; v1-rounds/ is
gitignored internal machinery and the gate itself is not in README either. CLAUDE.md was emptied
of exactly this kind of row in session 130 and is not the place to put one back.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 06:09:12 +02:00
..
fase1-spikes fix(fase1): spike B fan-out measures real conversation bleed, not a counter 2026-06-24 11:09:55 +02:00
plan docs(f16): the 1.18.0 assessment measured -- 0 of 19 forms moved, three features assessed and none built 2026-09-12 17:18:46 +02:00
rapport docs: fjern brukernavn, hjemmekatalog-stier og privat namespace fra publisert flate 2026-08-18 13:20:23 +02:00
research docs(research): MAF 1.9.0 capability map — feature-utilization for Fase 2 [skip-docs] 2026-06-24 11:36:26 +02:00
2026-06-24-two-approaches-brief.md docs: plain-text brief — goal + two approaches (MAF vs Claude Agent SDK) + learning goal 2026-06-24 09:21:31 +02:00
2026-06-26-fot-i-bakken.md docs(fot-i-bakken): ground-truth-verifisert levert-vs-lovet — agentiske lag inerte 2026-06-26 15:25:03 +02:00
2026-07-15-foundry-auth-recipe.md docs(1b): Claude på Foundry er en TREDJE klientflate — FoundryChatClient kan ikke binde den 2026-08-13 20:31:36 +02:00
2026-08-13-demo-presentasjon.html docs(deck): begge deck bærer det MÅLTE 1b-utfallet — ærlighet, ikke seier 2026-08-14 20:28:35 +02:00
2026-08-13-fase1a-lokal-ende-til-ende.md docs(fase1a): «uansett modell» hvilte på en måling jeg aldri leste ferdig 2026-08-13 20:19:31 +02:00
2026-08-13-fase4-azure-yaml-valg.md feat(4b): AZURE-profilen leser miljøet sitt, ikke operatørens laptop 2026-08-13 22:27:21 +02:00
2026-08-13-fase4-research-spike.md docs(4·): containeren er bygget — og wheelen er ikke installerbar alene 2026-08-13 21:57:40 +02:00
2026-08-13-presentasjon-ledelse.html docs(deck): tredje deck bærer samme MÅLTE 1b-utfall som de to andre 2026-08-18 09:37:14 +02:00
2026-08-14-fase1b-forste-levende-kjoring.md docs(1b): koordinatene byttet mot plassholdere - vei A, redigert framover 2026-08-18 13:03:11 +02:00
2026-08-18-vurdering-azure-omdoeping.md feat(gate): pakke-gaten leser INNHOLDET, ikke bare filnavn + vurdering av Azure-omdøping 2026-08-18 13:42:02 +02:00
2026-08-25-fable-misjonsreview.md docs: Fable 5 misjons-review — nærmer systemet seg faktisk målet? (ORDRE 20260825T104711Z) 2026-08-25 14:12:23 +02:00
2026-08-25-syretest-vei-ab.md feat(explore): katalogkallet koster O(baser), ikke O(korpus) (ORDRE 20260825T213645Z) 2026-08-26 14:45:16 +02:00
2026-08-26-katalogkostnaden.md feat(explore): katalogkallet koster O(baser), ikke O(korpus) (ORDRE 20260825T213645Z) 2026-08-26 14:45:16 +02:00
2026-08-29-maf-gjelden-omfang.md docs: MAF-gjeldens omfang malt, ikke bygget (F3/F15/F16/U16-17-19, ORDRE 20260825T214801Z) 2026-08-29 09:30:01 +02:00
2026-09-02-f15-maf-pinnen.md docs(f15): nevner-disiplin anvendt paa mitt EGET instrument - 17/16/2, ikke 16/16/2 2026-09-02 19:47:11 +02:00
2026-09-02-misjonsreview-v2.md docs: misjonsreview v2 mot bruksscenarioet — målt, ikke bygget (S2, ORDRE 20260902T113744Z-1245330375) 2026-09-02 17:17:39 +02:00
2026-09-02-read-bundle-kontekstkostnad.md docs: name the mechanism behind the untouched-debate proof, and re-measure all three bases 2026-09-03 01:30:09 +02:00
2026-09-03-forslag-fra-mandat.md docs(s7b): maaledokumentet baerer funnet CLAUDE.md-raden peker paa 2026-09-03 22:40:56 +02:00
2026-09-03-hierarkisk-navigasjon-k2.md docs(s7a-3): ordrens ANDRE binding maalt paa alle 478 K2-nivaaer, og to prosa-hull lukket 2026-09-03 08:09:53 +02:00
2026-09-03-syretest-s7a-k2.md docs(s7a): SS 5s "deler aarsak" var en paastand, ikke en maaling - og maalingen felte tallet 2026-09-03 03:42:18 +02:00
2026-09-03-syretest-s7a2-k2.md fix(docs): rediger vekk absolutt hjemmesti i S7a-2-rapporten - handover-gaten var roed paa HEAD 2026-09-03 06:02:17 +02:00
2026-09-04-major2-proposal-review-k2.md docs(major2): steg 9 fullfoert - M38..M40 roede, K2-omkjoeringen maalt, last_ruling-paastanden rettet 2026-09-05 20:39:52 +02:00
2026-09-04-s2c-debatt-k2.md docs(verdict-gate): maaleavsnittet og invariantraden - fem mutasjoner, fire tall verifisert 2026-09-04 20:41:32 +02:00
2026-09-04-syretest-s7b-k2.md docs(s7b): syretesten paa K2 - maaledokumentet med de fire tallene og tretten roede 2026-09-04 09:04:51 +02:00
2026-09-05-major2-trekreview.md docs(major2): remedieringsfoot under den sporede trekreview-kopien 2026-09-05 22:31:06 +02:00
2026-09-06-major2-levende-k2.md docs(major2): funn (b) re-maalt live - loekka konvergerer paa foerste reviderte forsoek 2026-09-07 00:52:39 +02:00
2026-09-07-okf-prepass-i-debatten.md docs(prepass): before/after on K2, the mutation battery and the invariant row 2026-09-07 14:39:50 +02:00
2026-09-07-prepass-mater-q5b-k2.md feat(prepass): --prepass-seed makes the cut a starting point, and K2 says it costs 2026-09-08 04:02:09 +02:00
2026-09-07-syretest-s7-prepass-k2.md docs(s7): a live model reads a declared cut -- three premises felled, four findings [skip-docs] 2026-09-07 16:07:29 +02:00
2026-09-08-f3-f4-nekten-og-forankringen.md feat(navigation,validator): read_dir names the rung that reads a document, and a run can require its anchoring 2026-09-08 05:32:46 +02:00
2026-09-08-funn-4-5-og-read-nekt.md feat(prepass,explore): padding dies at the prompt; a refusal the model can act on is a return value [skip-docs] 2026-09-08 23:36:54 +02:00
2026-09-08-funn-99-chatclient-refusal.md fix(cli): a provider failure leaves the CLI as one line, and a guessed base id is correctable 2026-09-08 21:03:18 +02:00
2026-09-08-n-bundlene-hypoteseform.md docs(n-bundlene): P4 -- one free N100 run after the P3 fix, (b') is now yes 2026-09-08 17:31:07 +02:00
2026-09-08-n-bundlene-konsum.md docs(n-bundlene): use ~-form for two absolute home paths -- handover gate was red on HEAD [skip-docs] 2026-09-08 15:23:55 +02:00
2026-09-08-syretest-s7c-begge-laaser-k2.md docs(s7c): both locks open, the price is delivered -- and the live model still did not read it [skip-docs] 2026-09-08 07:00:47 +02:00
2026-09-09-p6-betalt-k2-etter-kollaps.md docs(p6): one paid K2 run says the collapse bought legibility, not use 2026-09-09 13:14:55 +02:00
2026-09-09-p7-forankrede-identifikatorer.md feat(validator,generate,run): an identifier a proposal builds on must be in the input, or the verdict falls [skip-docs] 2026-09-09 16:04:01 +02:00
2026-09-09-p8-forankringstilbudet.md feat(run,generate): a run says what its delivered input can ground, before it spends an attempt [skip-docs] 2026-09-09 17:50:02 +02:00
2026-09-10-p9-v070-bundler.md docs(p9): okf v0.7.0 measured - the rename costs nothing, content adds 15 identifiers [skip-docs] 2026-09-10 06:33:57 +02:00
2026-09-10-p10-konform-k2-payload.md docs(p11): okf 0.8.1 measured - source-quota carries the denominators, one prior constant ranked the price schedule out [skip-docs] 2026-09-11 01:38:58 +02:00
2026-09-11-p11-okf-081.md docs(p11): okf 0.8.1 measured - source-quota carries the denominators, one prior constant ranked the price schedule out [skip-docs] 2026-09-11 01:38:58 +02:00
2026-09-11-p12-maf-gjeld.md docs(p12): MAF debt re-measured U1-U19 - 6/5/8 unchanged row for row, the pin bump moved U5's premise not its status [skip-docs] 2026-09-11 10:50:14 +02:00
2026-09-12-f16-maf-1180.md docs(f16): the 1.18.0 assessment measured -- 0 of 19 forms moved, three features assessed and none built 2026-09-12 17:18:46 +02:00
2026-09-12-p13-okf-pin-r761.md test(p13): the okf pin is MEASURED and deliberately NOT lifted -- the V1 stamp form disarms the forgery guard 2026-09-12 21:41:25 +02:00
2026-09-12-p13b-okf-bump.md fix(okf): close the inert ingest-stamp guard, then land okf 0.8.5 -- and read the block sources form all four bases actually write 2026-09-12 23:02:38 +02:00
2026-09-12-p14-kontekstsett.md fix(okf): top-level frontmatter title survives a nested sources: title 2026-09-13 07:40:57 +02:00
2026-09-14-p16-stressrunde-1.md docs(p16): stressrunde 1 -- four paid runs, four verdicts, the good/bad/ugly 2026-09-14 12:47:09 +02:00
2026-09-14-p18-stressrunde-2.md feat(p19): the trace says HOW, and a run says what it spent and why it stopped 2026-09-15 03:05:52 +02:00
2026-09-15-p17b-across-bundles.md docs(p17b): one commission across two bases -- the paid run, and the arm that got through 2026-09-15 05:05:47 +02:00
2026-09-15-p19-stressrunde-3.md docs(p19): stress round 3 -- five paid runs, and the premise the order rested on 2026-09-15 03:31:34 +02:00
2026-09-15-p20-stressrunde-4.md docs(p20): stress round 4 -- five paid runs, and the order's own rule felled by measurement 2026-09-15 08:28:54 +02:00
2026-09-15-p21-stressrunde-5.md docs(p21): stress round 5 -- the first ANCHORED round, and what the anchoring cost 2026-09-15 12:08:51 +02:00
2026-09-16-p22-stressrunde-6.md docs(p22): stress round 6 -- the three findings closed, and what the closing exposed 2026-09-16 01:50:40 +02:00
bestille-en-kjoring.md feat(row6): a proposal whose approach declared no requirement is unsupported 2026-09-17 16:40:54 +02:00
ekspert-svar.md feat(hitl): ekspertdommen kan ikke oppstaa av stillhet (F2, ORDRE 20260825T214801Z) 2026-08-27 01:22:07 +02:00
extending.md docs: llms.txt + sikkerhetskontakt til security@ (ORDRE 20260821T041218Z) 2026-08-21 11:29:54 +02:00
invarianter.md feat(round-builder): one command turns a run's outbox into a round the gate can read [skip-docs] 2026-09-19 06:09:12 +02:00
knowledge-base-recipe.md docs: kunnskapsbase for ÉN konkret kjøring — kategorier, innholdstyper, veiprosjekt-eksempel (ORDRE 20260821T083046Z) 2026-08-21 11:10:59 +02:00
kort-presentasjon.html docs(deck): begge deck bærer det MÅLTE 1b-utfallet — ærlighet, ikke seier 2026-08-14 20:28:35 +02:00
kunnskapsbase-for-en-kjoring.md feat(visibility): en lenke som ikke ble fulgt sier det - spor + betinget linje (ORDRE 20260821T142704Z) 2026-08-21 17:18:18 +02:00
okf-konsum-kontrakter.md docs(s7a-3): ordrens ANDRE binding maalt paa alle 478 K2-nivaaer, og to prosa-hull lukket 2026-09-03 08:09:53 +02:00
presentasjon-bygge-kunnskapsbase.html feat(visibility): en lenke som ikke ble fulgt sier det - spor + betinget linje (ORDRE 20260821T142704Z) 2026-08-21 17:18:18 +02:00
presentasjon-fagpersonens-bidrag.html docs: fagpersonens bidrag - spørsmål/sjekkliste til fagperson (ORDRE 20260824T092912Z) [skip-docs] 2026-08-24 15:24:51 +02:00
presentasjon-teknisk-arkitektur.html docs(deck): Del 1-3a rebuilt on verified seams — MAF-free guard named exactly, backends chain, five startup contracts, three debate stops 2026-09-18 07:11:31 +02:00
review-2026-07.md docs(review): kryssmodell-review 2026-07 (14 funn, 11 detach-bevis) + revidert roadmap + sesjonsplan Fase 2-6 2026-07-10 06:28:11 +02:00