Changelog

  1. The workspace now opens on where to start

    The Gandhi 2.0 workspace used to make you scroll through a problem, a change and a step before you reached an action. It now opens on the goal, the action to start with and why that action comes first: nothing recorded has to happen before it, and the other seven actions come after it. Four lines underneath say how the action serves the goal, and each opens in place when you want to check it. The other views follow the same rule: question pages list their parts by title, and a statement page shows the statement before what leads to it. The record holds no judgement of which action matters most, so the page reads the order from the record's own connections and says so. The TDI workspace and the private view the skill opens work the same way. Still untested with readers.

  2. A faster, draft-first version of the skill

    There is now an experimental ltp-v2 skill alongside ltp. Instead of building up a goal one question at a time, it drafts it straight away from your notes or a short description, then works through problems, conflicts and actions in a fixed order. It critiques its drafts before showing them, pushes back on your edits, shows trees as plain text that fits a chat window, and keeps a readable summary next to its saved state.

    The same brief, first reply: the original ltp opens with an introduction and a question; ltp-v2 opens with a draft goal tree and a short critique

    Install it with npx skills add reasoncommons/reasoncommons -s ltp-v2 (keep ltp installed too) and call it with /ltp-v2. The skill itself is skills/ltp-v2/SKILL.md. The original ltp is unchanged. For how the two compare on the same problem, see Drafting first vs asking first.

  3. A second way to read Gandhi 2.0

    The Gandhi 2.0 workspace is a prototype that reads the same record as the Gandhi 2.0 commons in a different order. It opens on one path from the goal to a first action, and any stop on it can be swapped for another the record links there. From any statement you can trace what leads to it and what follows, and a side panel shows its exact wording, who proposed it, its sources and what is still open. The commons pages are unchanged. We have not yet tested it with readers.

  4. Follow a complete Adaptive Persistence engagement at Northbank Print

    Northbank Print now follows a complete fictional Adaptive Persistence engagement, from agreeing the issue through assessment, capacity choice, the ACT group conversation and repeated review. Short illustrated views link to dated working records, retain the original results and corrections, and leave the next review visibly pending.

  5. Share your commons on the site

    A group using the skill can now ask it to share their commons on the commons page. The group chooses what goes public, everyone named agrees to the name they choose, and they see a preview first. The skill then opens a pull request for a maintainer to review. Shared commons appear on their own shelf under Creative Commons Attribution 4.0, marked as shared by their participants. You need a GitHub account; the skill guide explains each step.

  6. See your commons in the browser

    Ask the skill to "show me our commons in the browser" and it lays out the whole working file as pages like the demonstration commons: an overview, a map, a page for each question, and every statement with who contributed it, who accepted it and what is still open. Accepted, proposed and disputed reasoning are all shown, each marked. The pages stay on your computer, beside the working file, and nothing is published or sent. See the skill for how to ask.

  7. Corruption in Kenya, 2014 joins the demonstration commons

    Corruption in Kenya, 2014 asks why senior officials so rarely answered for corruption. It draws on Transparency International’s Anti-Corruption Helpdesk brief of October 2014: bribery, school funds lost and billions spent without records, a commission that could not prosecute, and two dilemmas over recovering stolen money. Every statement has a page reference and describes Kenya as of 2014; the AI reconstruction has no endorsement by Transparency International or recorded trial results.

  8. Our own plan in the demo now matches the restructured record

    The prerequisites and actions pages of our own commons now show the plan as restructured on 26 October. The actions keep only steps whose effect could be disputed, each with a need and an expected effect, and end at the Stage 1 goal; plain tasks such as grant submissions now live in our task tracker. Everything on both pages is still proposed.

  9. Goal trees read top to bottom, one level at a time

    Goal trees in Reason Commons in use now read from the top down, with each condition indented beneath what it is needed for, so a large tree no longer runs off the side of the screen. Switch on Step through to start from the goal alone and open one level at a time: try it on Reason Commons' own goal. It starts switched on for larger trees and remembers your choice.

  10. Our own commons now shows the reworked reasoning

    The Reason Commons commons now draws on Rufus and David's October rework of their reasoning: 63 statements and 57 connections that Rufus has accepted for now, across the goal, what stands in the way, four dilemmas and the changes chosen in response. Reasoning about funding and their own working arrangements stays in the private working record, and proposed prerequisites and actions will appear once they are accepted.

Built with LogoFlowershow