00 — The Brief: wardley-maps.sgit.ai

Version v0.33.62 · 23 August 2026 From Dinis Cruz, via the SG/Send Librarian To the agent commissioned to build wardley-maps.sgit.ai Licence CC BY 4.0 — but read 07__boundaries-and-licensing.md before you copy anything, because this site sits inside a CC BY-SA ecosystem and that is the single biggest trap in the whole build

1. The commission

Three jobs, in the words they were given:

1. Consolidate "what I have created on Wardley maps on the __Send site and the multiple *.sgit.ai sites" — including "one of them in addition to having a page and vault with maps, we even mapped the doctrines".
2. Job A — "do a research on other materials (videos, infographics, docs) that I have published around wardley maps".
3. Job B — "find the best industry resources about it, and per resource create a page with the details, screenshots and links to those resources".

All three are done. Jobs A and B are 03__ and 04__. The consolidation is 01__, 02__ and 05__.

2. Two corrections to the premise, up front

The doctrine mapping is not in a vault. It is static HTML on pki.sgit.ai, at /packs/registry-mvp/doctrine.html, with its raw data at /packs/registry-mvp/doctrine/doctrine.json — 41 records: 15 strong, 11 partial, 6 weak, 9 no-basis, across Wardley's 4 phases and 6 categories. pki.sgit.ai/admin/index.html states the site is "not served from an sgit vault". The site that has a page plus a vault of maps is sgit.ai (the Strategy Maps vault, ookq4mn4, 33 files, 830 KB), and that vault has no doctrine content at all. So the thing remembered as one property is actually two, and the doctrine work is currently buried at appendix depth inside a PKI pack. That single asset is the strongest reason to build this site, and moving it is task one.

There is no public bridge between the security-era Wardley work and the agentic-era Wardley work. Job A found 26 published items running 2018→2025. Job A also found that the 2025–2026 agentic framing — the Explorer/Villager/Town Planner agent teams, de-commoditisation, the skill lifecycle — has never been published anywhere. The only public link between the two eras is one clause in one co-authored MCP article from June 2025. Do not write the site as though that bridge already exists in public. Building it is the site.

3. The thesis

Maps are claims, not pictures — and almost nobody publishes theirs where they can be argued with.

The corpus already says this out loud: "A Wardley map is a claim, not a picture: it asserts where each component sits on the evolution axis and is therefore arguable" (14 Aug 2026). Everything the site does follows from taking that seriously — maps in git, coordinates in text, evidence attached to placements, and a doctrine assessment that names the artefact behind every rating.

There is a second, sharper thesis available, and the research supports it hard:

The Wardley ecosystem's link layer has rotted, and its own canonical indexes are now wrong.

As of August 2026: MapKeep is dead (ceased 30 May 2026 — the only tool with real-time collaboration), MapScript is dead, every doctrine assessment tool is dead, the Map Camp Slack invite 404s, and the Leading Edge Forum course 404s. All five are still listed as live by awesome-wardley-maps, by Wikipedia, and in two cases by Simon Wardley's own resources page. Two Map Camp domains contradict each other about whether the conference still exists. A site that simply verifies every link and publishes the verification date is immediately more useful than the community's own index. That is a low bar and a real service.

4. What is original here — the five pages only this site can write

Ranked. These are the reason to build it rather than curate someone else's list.

Two more that carry a paragraph each, not a page: the commodity illusion ("a lot of things that feel like a commodity or a product to the user are actually very weak and very immature behind the scenes") and the agent as a user.

And the most quotable line in the whole corpus, which belongs on the front page or nowhere:

"Anybody who spends a lot of money on tokens has an engineering problem. They are using explorer-type code and solutions in a commodity environment." — 4 June 2026

5. The honesty constraint

Every sibling site publishes a /shipped/ page separating what exists from what is designed. This site inherits it, and here it has real work to do, because the corpus over-promises in a specific and documentable way:

• Zero of the 13 Mermaid map sources in the repo currently render. Every one has the wrong code fence — bare ` or `wardley-beta instead of `mermaid — and one note line is missing its required quotes. The 8 PNGs from 24 May are the only working visuals in the entire corpus.
• The living evolution map was abandoned on 26 February 2026 — six months stale, against the corpus's own warning that "a map that was accurate three months ago and hasn't been updated is worse than no map."
• No doctrine assessment and no gameplay analysis was ever produced inside __Send, despite both being specified as standing Cartographer responsibilities. (The one real doctrine assessment that exists is the pki.sgit.ai one — a different project.)
• 31 of 33 ROLE.md files never mention Wardley. The Explorer Cartographer's own 2,057-word role definition — the role the founding brief called "central to this entire structure" — mentions Wardley zero times.
• "Maps are graphs with position" has no implementation. It is architecture; the reality doc lists it as PROPOSED.
• No .owm files, no map JSON, no map SVG existed in the corpus before this pack.

Say all of that plainly. The gap between the specification and the artefacts is itself an honest and interesting Wardley story — a map of the mapping practice would put most of it at Genesis.

6. What this pack ships that is new

Not just analysis — four working artefacts:

7. Why the screenshots are a script, not a folder of PNGs

Job B asked for screenshots per resource. This pack ships the capture kit rather than the images, and the reason is worth stating plainly rather than hiding: the sandbox this pack was assembled in has an egress proxy that allows only package registries. Every third-party host returned ERR_TUNNEL_CONNECTION_FAILED — including sgit.ai itself. Nothing was captured, and I did not route around the block.

So screenshots/capture.js does the job wherever the network is open — npm i playwright && node capture.js. It writes captured.json recording the HTTP status, page title, final URL after redirects, and a UTC timestamp for every attempt including the failures. Publish that file next to the images. Given §3 — an ecosystem whose own indexes point at dead tools — a screenshot with a provable capture date is worth considerably more than a screenshot without one.

8. The numbers

9. Build order

1. Move the doctrine assessment. /doctrine/ — the 41 records, the JSON endpoint, the phase filter, the evidence-per-rating discipline. Leave a stub on pki.sgit.ai. This is the site's anchor asset and it already exists.
2. Fix the 13 broken map fences and render them. Ship the SVGs. Use the pipeline in 05__ §4 — it is proven, and this pack's four maps are the proof.
3. /resources/ — Job B. 34 pages from 04__ and sources__industry-resources.json. Run capture.js first so every page opens with a dated screenshot. Lead with the verified-in-August-2026 angle.
4. /dinis/ — Job A. The chronology from 03__, embedding the 13 YouTube recordings. Then archive the fragile ones — the SlideShare deck is the only artefact of the Map Camp talk and it is behind a Cloudflare wall.
5. /patterns/pst/ — the Explorer/Villager/Town Planner write-up. The site's most original page. 02__ is the raw material; it has never been written up as a pattern anywhere.
6. /method/ — de-commoditisation, the custom-axis verdict, the broken middle, maps-as-graphs. Four opinionated pages.
7. /agents/ — the machine surface: the coordinate contract, the parse rules, the definitions endpoint, and the epistemic guardrail from 04__ (an agent placing components on the evolution axis is generating consensus-shaped output; it must surface placements as contestable claims, never findings).
8. /shipped/ — §5, in full, without softening.

Publish the open questions and the tensions unresolved, per the house pattern. 08__ supplies both.

This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0).