open/okr
1
0
Fork 0
Code Issues Releases 3 Activity

fix(okr): C6 markedsoppdatering + kaskade F-i (verifisert-bruker, align-not-cascade, NCT)

Browse source 
C6 markedsclaim -> verifisert-bruker-disiplin (README):
Paastanden "increasingly by Norwegian public sector organizations like NAV and
FINN.no" hadde ingen kilde i kunnskapsbasen. Erstattet med fire adoptorer som
ALLE har en offentlig kilde i okr-sources.md par. 4 (Digdir, NAV-team, Oslo
Origo, FINN.no) + eksplisitt setning om at usourcede virksomheter ikke navngis.
Skatteetaten/Entur/Politiet er fortsatt IKKE verifisert og navngis derfor ikke.

okr-sources.md par. 4: NAV manglet helt som oppslag selv om README paastod
bruken. Lagt til med first-party-kilde (aksel.nav.no produktbloggen), scopet
til team-/produktnivaa - etatsnivaa-OKR er ikke dokumentert.

okr-sources.md par. 7 (ny): alternative rammeverk med verifisert attribusjon.
- NCT (Narrative, Commitments, Tasks) tilskrives Ravi Mehta / Reforge, med
  eksplisitt advarsel om den vanlige feilattribusjonen til Radical Focus 2. utg.
- Evidence-Based Management -> 2024-guiden. Kun det verifiserte "what's new"
  er gjengitt (KVA-til-maaletype-kobling, Input/Impact som egne maaletyper,
  klargjorte KVA-beskrivelser). Mission/Vision-innrammingen fra raarapporten
  er UTELATT - ikke bekreftet av Scrum.orgs egen What's-New.
Sist oppdatert-markoer bumpet Januar -> Juli 2026.

align, don't cascade (kaskade.md + kaskadebygger-agent.md):
Begge flater laerte mekanisk kaskadering ("et overordnet KR blir teamets
Objective"). Rammet om til alignment via lineage: org-KR er INPUT til teamets
Objective, teamet omformulerer og eier formuleringen, team-KR maaler teamets
eget bidrag, og manglende paavirkbarhet rapporteres som gap i stedet for et
konstruert bidrag. Individ-OKR eksplisitt utelukket paa begge flater.
NAV-praksisen er brukt som norsk anker for doktrinen.

F-i (kaskade.md): siste gjenstaaende foreldede kontekstblokk skrevet om til
post-1.6.0-moensteret (analyse.md:13-20). Glob lagt til i allowed-tools siden
blokken naa instruerer Glob.

Planavvik (premiss-verifisering): Step 13 sa README "16" -> "17 domenefiler".
Ground truth motsier premisset - freshen-references.md scorer 16 AV 17 filer
(okr-quality-rubrics.md er eksplisitt ekskludert som sirkulaer). "17" ville
gjort README feil. Skrevet presist i stedet: "16 of the 17 ... the quality
rubric itself is excluded". Planens Verify (grep -c "16 domain" -> 0) passerer.

Vakt-case (i) GROENN. Alle 13 konsistens-cases groenne; suite 163/163/0.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016pUfkQ1YyH75z9y6RBaBHG
This commit is contained in:
Kjell Tore Guttormsen 2026-07-25 06:32:18 +02:00
commit 1cb2eed5a2
4 changed files with 85 additions and 21 deletions
Download patch file Download diff file Expand all files Collapse all files

4
README.md
View file

@ -20,7 +20,7 @@
Every organization has a strategy. Few manage to turn it into goals that teams actually work toward.
OKR (Objectives and Key Results) is a proven framework for that translation — used by Google, Intel, and increasingly by Norwegian public sector organizations like NAV and FINN.no. But adopting OKR is hard. The methodology sounds simple ("write inspiring goals with measurable results") until you try it. Then you hit real questions:
OKR (Objectives and Key Results) is a proven framework for that translation — used by Google and Intel, and documented in Norwegian practice: Digdir runs OKR in its product delivery model, NAV product teams use it at team level, Oslo Origo built and open-sourced its own OKR tracker, and FINN.no has more than six years of it. Every adopter named here has a public source in `skills/okr-offentlig-sektor/references/okr-sources.md` § 4; organizations we could not source are not named. But adopting OKR is hard. The methodology sounds simple ("write inspiring goals with measurable results") until you try it. Then you hit real questions:
- *How do we connect our OKR to the goals in our tildelingsbrev?*
- *What's a good Key Result vs. just an activity disguised as one?*
@ -147,7 +147,7 @@ Drop documents — tildelingsbrev PDFs, virksomhetsplan docx, meeting notes —
> /okr:freshen-references
```
`/okr:help` lists all 14 commands, 7 agents, and a recommended workflow keyed to where you are in the tertial cycle. `/okr:freshen-references` is a knowledge-base self-evaluator: it scores the 16 domain reference files against an anchored quality rubric and polls named public sources (UN EGDI, EU eGov Benchmark, OECD DGI, Digdir, WCAG, forvaltningsloven) to flag outdated `Sist oppdatert` markers.
`/okr:help` lists all 14 commands, 7 agents, and a recommended workflow keyed to where you are in the tertial cycle. `/okr:freshen-references` is a knowledge-base self-evaluator: it scores 16 of the 17 domain reference files against an anchored quality rubric — the quality rubric itself is excluded, since scoring the scoring instrument is circular — and polls named public sources (UN EGDI, EU eGov Benchmark, OECD DGI, Digdir, WCAG, forvaltningsloven) to flag outdated `Sist oppdatert` markers.
---