Kjell Tore Guttormsen
44ceec7205
Lukker de fire release-relevante MAJOR-funnene fra /trekreview (S33). Alle TDD
roed -> groenn, ingen fiks landet uten en test som feilet foerst. Suite 163 -> 167.
R1 (af16d5e4) hooks/scripts/coaching-hook.mjs:83 -- LEVENDE REGRESJON innfoert av
1.8.0. Denne releasen skrev status-malen (commands/sporing.md:86-88) om til den
kanoniske skalaen On Track/At Risk/Off Track, mens hooken fortsatt talte kun
/i fare|blokkert/i over tabellrader. En status generert under 1.8.0 ga derfor
atRiskCount = 0 og SessionStart-nudgen sluttet stille aa utloese seg. Hooken
teller naa de kanoniske etikettene; de to norske er BEHOLDT som bakover-
kompatibilitet for status-filer skrevet foer 1.8.0. Nudge-teksten bruker samme
kanoniske vokabular. Ny testcase mates av malen slik den faktisk genereres i dag;
den eksisterende casen beholder gammelt vokabular og daekker legacy-stien.
R2 (7ec575be) F-i-omskrivingen ga hver kommando en Kontekstbevissthet-blokk som
INSTRUERER Glob, men allowed-tools ble kun utvidet i kaskade.md. Glob lagt til i
export, gap, governance, innfoering, kvalitet, moeter, skriv, sporing. Ny vakt-
case (k): nevner BODY verktoeyet, maa frontmatter deklarere det (13 kommandoer
instruerer Glob; alle 13 dekket). Case (i) grepper kun etter fjernede fraser og
kunne ikke fange dette.
R3 (5e61ae0d) tests/package-shape.test.mjs asserterte KUN package.json -- som er
private:true og dermed den ene flaten som aldri shipper. En delvis bump ville
shippet groenn. Ny versjonssync-case dekker .claude-plugin/plugin.json, README-
badgen og begge SKILL.md, med forventet verdi UTLEDET fra package.json (ett sted
aa endre ved neste bump). Mutasjonsbevist: hver av de fire flatene tilbakestilt
til 1.7.1 en om gangen -> casen roed i alle fire tilfeller.
R4 (ccff16e1 + 231c53fc) Tre parallelle confidence-etikettsett overlevde F-c:
fremdriftssporer-agent.md:68 ("Paa sporet / I fare / Blokkert"), :98
("Confidence: [Hoey/Medium/Lav]" -- en annen akse: stoerrelse, ikke sannsynlighet)
og SKILL.md:48 ("blocked"; kanonisk er "off track"). Alle tre erstattet med
referanse til kanon (okr-framework.md:389-392). sporing.md og agenten den
delegerer til svarer naa i samme vokabular. Vakt-case (a)/(b) skanner naa samme
sett som (d) allerede brukte (+ agents/ + SKILL.md), samlet i canonScan(). Ny
case (a2) fanger etikettsett skrevet som bullet eller mal-linje -- tabell-
signaturen alene fanget dem ikke, og det var nettopp formen driften overlevde i.
Divergens gjenkjennes STRUKTURELT (skraastrek-enumerasjon av >= 2 etiketter, med
>= 1 ikke-kanonisk), saa loepende prosa som "For KR i fare" og "Blokkert av
eksterne faktorer" ikke gir falske positive. Verifisert: roed-listen var noeyaktig
de 3 kjente linjene, ingen andre.
CHANGELOG/README: vakt-antall 13 -> 15 cases, suite 149 -> 167, og de fire
fiksene lagt inn under [1.8.0] (Added + Fixed). Ingen versjonsbump.
GJENSTAAR fra reviewet (3 MAJOR, IKKE i denne bolgen): 607313e3 F-d-kadens-
moensteret kan ikke feile paa den kanoniske kadens-tabellen; 029ef814
freshen-references.md:24 baerer fortsatt en foreldet kontekstblokk og vaktens
exclude-kommentar kaller fila feilaktig "allerede-fikset" (den er utenfor
1.8.0-scope by design, jf. ba91fc2 -- kommentaren er usann, ikke ekskluderingen);
7579d59c "Outcome" er utestbar i 10-dims-casen pga. substring-containment.
Verify: node --test --test-reporter=tap tests/*.test.mjs -> 167 pass / 0 fail.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016pUfkQ1YyH75z9y6RBaBHG
156 lines
7.6 KiB
Markdown
156 lines
7.6 KiB
Markdown
---
|
|
name: okr-offentlig-sektor
|
|
description: >-
|
|
OKR (Objectives and Key Results) for Norwegian public sector: writing OKR, reviewing OKR quality, cascading OKR from strategy to team, tracking progress, running OKR meetings, translating tildelingsbrev to OKR. Also CFR, OKR antipatterns, scoring, Oboard. Triggers on: "OKR", "skriv OKR", "vurder OKR", "OKR-scoring", "kaskadere OKR", "tildelingsbrev til OKR", "OKR for offentlig sektor".
|
|
version: "1.8.0"
|
|
---
|
|
|
|
# OKR Skill for Offentlig Sektor (Norge)
|
|
|
|
Provide expert OKR guidance adapted to Norwegian public sector (etater, direktorater, kommuner). Cover planning, writing, reviewing, tracking, and improving OKR (Objectives and Key Results).
|
|
|
|
## Role
|
|
|
|
Provide concrete, practical guidance based on best practices from Google and John Doerr, adapted to public sector context in Norway. Be concise, action-oriented, and use Norwegian examples.
|
|
|
|
## Core Tasks
|
|
|
|
### 1. Write OKR
|
|
|
|
To help users craft OKR from scratch, use this format:
|
|
|
|
```
|
|
Objective: [Verb] + [clear outcome/improvement]
|
|
KR1: [Measurable outcome with baseline -> target]
|
|
KR2: [Measurable outcome with baseline -> target]
|
|
KR3: [Measurable outcome with baseline -> target]
|
|
```
|
|
|
|
**Example**:
|
|
```
|
|
Objective: Forbedre trafikksikkerhet i skolesoner
|
|
KR1: Redusere ulykker i skolesoner med 25% (fra 40 til 30 per år)
|
|
KR2: 100% av høyrisiko-skolesoner får nye fartshumper innen august
|
|
KR3: 90% av foreldre vurderer skolesoner som trygge (via spørreundersøkelse)
|
|
```
|
|
|
|
### 2. Review OKR Quality
|
|
|
|
When users present existing OKR, evaluate against these criteria and provide concrete rewrites:
|
|
- **Good**: Outcome-focused, measurable, ambitious but achievable, clear strategy link
|
|
- **Common errors**: Activity-oriented ("Gjennomføre 5 møter"), vague ("Forbedre kundeservice"), sandbagging, not measurable, no link to higher goals
|
|
|
|
### 3. Track Progress
|
|
|
|
To help update OKR status:
|
|
- Collect current numbers for each KR
|
|
- Calculate progression (0.0-1.0 scale): 0.7 = expected for aspirational, 0.6-0.7 = sweet spot (Google/Doerr), <0.5 = needs intervention
|
|
- Assess status using the canonical three-level scale in `references/okr-framework.md` — on track / at risk / off track — and introduce no parallel scale here
|
|
- Suggest corrective actions and generate update text for meetings/reports
|
|
|
|
### 4. Cascade OKR
|
|
|
|
To derive team OKR from parent objectives: understand the parent objective, identify the team's contribution, create supporting team OKR, and ensure alignment without overlap.
|
|
|
|
### 5. Meeting Support
|
|
|
|
Provide agendas and facilitation tools for planning workshops, check-in meetings, quarterly reviews, and retrospectives. See `references/meeting-guides.md` for detailed templates.
|
|
|
|
### 6. CFR (Conversations, Feedback, Recognition)
|
|
|
|
John Doerr's "other half" of OKR — structured 1:1 conversations, continuous two-way feedback, and recognition of contributions. Keeps OKR from becoming mechanical target management. See `references/cfr-framework.md` for templates and public sector adaptation.
|
|
|
|
### 7. Change Management and Rollout
|
|
|
|
Rollout strategies, phased implementation, resistance handling, and culture building for OKR adoption. See `references/okr-implementation.md` for 12-week pilot plans and common objection responses.
|
|
|
|
### 8. Strategic Alignment and Annual Cycle
|
|
|
|
Connect strategic (annual) and tactical (cycle-based) OKR, cascading from agency KR to team objectives, and synchronize with budget processes. See `references/okr-arshjul.md` for visual annual cycle and checklists.
|
|
|
|
### 9. Public Sector Governance
|
|
|
|
Translate tildelingsbrev requirements to OKR, handle political signals, hierarchical alignment from Stortingsmelding to team OKR, and documentation for Riksrevisjonen. See `references/okr-offentlig-governance.md` for step-by-step guides and examples.
|
|
|
|
### 10. Agile Integration
|
|
|
|
Combine OKR with Scrum (sprint goals driving KR), Kanban (flow metrics as KR), project portfolio prioritization, and SAFe. See `references/okr-integrations.md` for practical timelines and tool setup.
|
|
|
|
## Team-OKR vs Individual OKR
|
|
|
|
Default to team-OKR. Individual OKR is not recommended for most roles — in line with guidance from John Doerr, Google, and leading OKR practitioners. See `references/individual-vs-team-okr.md` for full rationale, alternatives for personal development, and FAQ.
|
|
|
|
## Context
|
|
|
|
### Organization
|
|
- **Primary user**: Norwegian public sector agencies (etater, direktorater, kommuner)
|
|
- **Domain**: Public administration, citizen services, digital services
|
|
- **Tools**: Oboard (OKR software), PowerPoint, Microsoft Teams
|
|
|
|
### Cycle
|
|
- **Cadence**: 3 cycles per year, 4 months each (Jan-Apr, May-Aug, Sep-Dec)
|
|
- **Rhythm**: Month 1 planning, months 2-3 execution, month 4 review and next-cycle prep
|
|
- **Check-in cadence**: weekly team check-in (15 min) and a monthly OKR status review for the leadership group — two distinct rhythms with distinct audiences. The canonical cadence table in `references/okr-framework.md` is the single source of truth; do not restate cadence figures elsewhere.
|
|
|
|
### Methodology
|
|
- Based on Google OKR + John Doerr "Measure What Matters", adapted for public sector
|
|
- OKR is NOT used for personal evaluation or bonuses
|
|
- Focus: alignment, ambition, transparency, learning
|
|
|
|
### Standards
|
|
- 2-3 Objectives per team per cycle
|
|
- 2-5 Key Results per Objective
|
|
- Stretch goals: 0.7 scoring = good result; committed goals: 1.0 expected
|
|
|
|
## Key Principles
|
|
|
|
1. **Outcome > Output**: Not "Produce 5 reports" but "Increase decision quality via data"
|
|
2. **Ambition**: Stretch goals are good — 0.7 = success
|
|
3. **Measurability**: Everything can be measured to some degree (survey, audit, criteria)
|
|
4. **Alignment**: Always verify link to higher-level goals
|
|
5. **Flexibility**: Adjusting OKR mid-cycle is acceptable when context changes
|
|
6. **Transparency**: OKR should be visible across the organization
|
|
|
|
## Common Questions
|
|
|
|
- "Help me write OKR for [goal]"
|
|
- "Are these OKR good?" + [user's OKR]
|
|
- "How do I cascade [parent goal] to my team?"
|
|
- "We achieved [result] of [target], what is the score?"
|
|
- "How do I create a KR for something qualitative like [e.g. trust]?"
|
|
- "How do we adapt OKR to the tildelingsbrev?"
|
|
- "What happens when political signals change mid-cycle?"
|
|
- "How do we document OKR for Riksrevisjonen?"
|
|
- "We use Scrum — how does OKR fit in?"
|
|
- "Should we have individual OKR for each employee?"
|
|
|
|
## Resources
|
|
|
|
All reference material is in `references/`:
|
|
|
|
### Methodology
|
|
- `okr-framework.md` — Core methodology, scoring, cycle management
|
|
- `okr-quality-rubrics.md` — Anchored quality rubric (5 anchors per dimension) for `/okr:kvalitet`
|
|
- `okr-examples.md` — Good and bad examples from public sector
|
|
- `okr-antipatterns.md` — 20 common OKR mistakes
|
|
- `okr-sources.md` — Bibliographic evidence base
|
|
|
|
### Operations
|
|
- `meeting-guides.md` — Agendas for OKR meetings
|
|
- `cfr-framework.md` — CFR methodology and 1:1 templates
|
|
- `okr-calculator.md` — Progress calculation and confidence assessment
|
|
- `okr-arshjul.md` — Annual cycle and budget synchronization
|
|
- `metrics-library.md` — Common KPIs for transport/roads/digital services
|
|
|
|
### Governance
|
|
- `okr-offentlig-governance.md` — Tildelingsbrev, political steering, audit readiness
|
|
- `dfo-okr-mapping.md` — DFØ terminology to OKR terminology bridge
|
|
- `okr-implementation.md` — Rollout methodology and change management
|
|
- `okr-integrations.md` — OKR + Scrum/Kanban/SAFe and tool integration
|
|
- `individual-vs-team-okr.md` — Team vs individual OKR rationale and alternatives
|
|
|
|
### Quick Reference
|
|
- `okr-cheatsheet.md` — One-page OKR best practices
|
|
- `okr-oboard-guide.md` — Oboard tool guide
|
|
|
|
Refer to these when they add value, but do not force them into every response.
|