Fiskwise

Fisheries intelligence, shared

Scientific knowledge.
Working with AI.

Agentic AI tools for fisheries science.

Fiskwise brings together AI skills, scientific knowledge, and R tools to help researchers prepare data, build models, and communicate results — with the researcher directing and being responsible for the work.

Explore the knowledge packs Installation and update guide
Built onRtidyverseQuarto
THE FISKWISE ECOSYSTEM
SHARED COREFiskwise + wise-writing

Scientific workflows, code, writing and review

RtidyverseQuarto
ggOceanMaps · ggFishPlots
DATA ADD-ONBAITIMR Biotic expertisePRESENTATION ADD-ONwise-presentationsScientific stories and slides
METHOD ADD-ONS
Independent packs. A shared foundation.

Knowledge you can inspectDocumented methods and version-controlled skills

Workflows you can reproduceCode, results, and scientific narrative together

Decisions you remain responsible forResearchers direct, review, and evaluate the work

01 / THE KNOWLEDGE PACKS

Expertise for each stage
of the work.

Start with the shared core, then add data expertise, presentations or a method. Every pack can also be used independently.

Core pack

wise-writing

The shared scientific foundation

Structure manuscripts, data reports, cruise reports, assessments, advice sheets, and benchmarks with shared writing conventions and Quarto templates.

Explore documentation GitHub
Data add-on

BAIT

Work with Biotic data locally

Connect to local data, prepare survey catches, extract sales-note and logbook catch series, explore maps and life histories, and follow documented confidentiality and sampling rules.

How to use this pack Explore documentation GitHub
Method add-on

index-wise

Build and evaluate survey indices

Move from qualified survey observations to a first model, validation, sensitivity analysis, and an index report. Uses BAIT for Biotic data and wise-writing conventions.

How to use this pack GitHub
Method add-on

spict-wise

Document a stock assessment

Prepare assessment inputs, fit and diagnose SPiCT models, explore alternatives, and produce reports and advice using wise-writing.

How to use this pack GitHub
Method add-on

gadget-wise

Build gadget3 models

Scaffold a gadget3 model in the IMR or Icelandic folder format, compile data through BAIT, mfdb or your own files, build, fit and reweight the model, run jitter, retro and leave-out diagnostics, project reference points and catch options, and write the chapter with wise-writing.

How to use this pack GitHub
In development

Knowledge packs are instructions, methods, practical tools, and review procedures that agents read at runtime. Any local coding agent with file and shell access can use them; they are not trained models.

02 / SHARED R TOOLS

R tools for
fisheries science.

Fiskwise checks ggOceanMaps and ggFishPlots during runtime setup and updates them with your package permission. They also work independently; use their documentation for mapping and life-history methods.

RstoxUtils

Read Norwegian commercial catch data: sales notes, ERS logbooks and FDIR codes. Optional; installed only when catch data are needed.

Explore documentation GitHub

ggSurveyPlots Planned

Visualization of fisheries survey data.

03 / GET STARTED

Start with a question.
Build on shared expertise.

Begin with wise-writing and the shared R tools. Add presentations, data and methods when needed. Fiskwise checks the shared R toolchain and packages with permission; add specialist data and modelling packs when needed.

  1. 01

    Install the shared foundation

    Give your agent the short install prompt. It checks what is available and explains any software needed.

  2. 02

    Describe your task

    Ask for a report, a survey-data summary, an index or a presentation. Your agent uses the relevant pack and helps add it when needed.

  3. 03

    Direct the work and review the result

    Set the scientific question, confirm assumptions, inspect code and diagnostics, and evaluate the conclusions. Agent output requires human review.

WHY FISKWISE?

Domain knowledge,
made reusable.

The name combines FISK — Fisheries Intelligence Skill Kit — with wise, reflecting the domain knowledge and understanding the ecosystem gives to an AI agent. It also draws on fisk, the Scandinavian word for fish, and fiskeri, meaning fishing or fisheries.

Developed at the Institute of Marine Research, Norway, Fiskwise includes tools for IMR data formats alongside knowledge and software useful to the wider scientific community.

As methods and workflows improve, version-controlled resources let those improvements travel between projects, researchers, and AI agents.

Help improve the ecosystem