The Librarian keeps a list of concepts that recur across the network, each with a definition, the page where it is defined, and the terms that signal it (data/concepts.json). The build searches every snapshotted page for those terms and lists, per concept, the pages and sites that use it: that list is on the concepts page, with every page linked. These mindmaps draw that list.
Two parts of this map are the Cartographer's and can be argued with. The seven themes are a grouping made for this map: no source groups the concepts this way. The site lists are what the build found on 25 September 2026: a site is under a concept when at least one of its snapshotted pages matches the concept's terms. A mention counts the same as a definition, so a site on a leaf uses the word; it does not necessarily build the thing. The concept pages give the counts, which change with every build, so the maps do not repeat them.
Site names ending in .sgit.ai are shortened: abp is abp.sgit.ai, llms is llms.sgit.ai, and so on. sgit.ai and riskmandate.ai keep their full names.
The seven themes
mindmap
root((Recurring concepts))
Keys and trust
[Read keys]
[Client-side encryption]
[PKI]
[Vault key management]
Vaults at work
[Vault apps]
[Append lanes]
[Telemetry]
[Twins]
[Fractal semantic graphs]
[Performance and cost]
Agent permissions
[Agent Behaviour Policy]
[Grants and mandates]
[Blast radius]
[Undo classes]
Risk and insurance
[Risk acceptance]
[Acceptance intervals]
[Licence to operate]
[Insurability]
[Security audits]
Standards and method
[EU AI Act]
[GDPR]
[Wardley maps]
[Open source]
Working in public
[Briefs]
[Interview pages]
[Lessons learned]
[Em-dashes and style]
[Synthetic users]
[Provenance]
[Frozen, hashed sources]
The business
[Business plans]
[Partnerships]
[Pricing ladders]
[Startups]
[Brand assets]Keys and trust
mindmap
root((Keys and trust))
[Read keys]
abp
elevenlabs.providers
games
graphs
llms
nhi
open-source
riskmandate.ai
risks
sgit.ai
standards
store
ungovr.providers
wardley-maps
what-can-it-do.games
[Client-side encryption]
abp
nhi
open-source
riskmandate.ai
sg-sentinel
sgit.ai
store
subscriptions
teams
what-can-it-do.games
[PKI]
games
llms
nfrs
nhi
open-source
pki
riskmandate.ai
sgit.ai
twins
wardley-maps
[Vault key management]
sgit.aiVaults at work
mindmap
root((Vaults at work))
[Vault apps]
abp
elevenlabs.providers
games
llms
pki
providers
riskmandate.ai
risks
sgit.ai
skills
store
ungovr.providers
[Append lanes]
abp
games
pki
riskmandate.ai
sgit.ai
what-can-it-do.games
[Telemetry]
abp
games
nhi
open-source
pki
riskmandate.ai
sgit.ai
what-can-it-do.games
[Twins]
abp
nfrs
pki
riskmandate.ai
sgit.ai
twins
[Fractal semantic graphs]
abp
graphs
sg-sentinel
sgit.ai
[Performance and cost]
sgit.aiAgent permissions
mindmap
root((Agent permissions))
[Agent Behaviour Policy]
abp
riskmandate.ai
sgit.ai
store
[Grants and mandates]
abp
elevenlabs.providers
games
pki
riskmandate.ai
sgit.ai
store
ungovr.providers
what-can-it-do.games
[Blast radius]
abp
elevenlabs.providers
graphs
llms
nhi
pki
providers
riskmandate.ai
risks
sgit.ai
ungovr.providers
[Undo classes]
abp
elevenlabs.providers
open-source
pki
providers
riskmandate.ai
risks
sgit.ai
ungovr.providers
what-can-it-do.gamesRisk and insurance
mindmap
root((Risk and insurance))
[Risk acceptance]
abp
games
graphs
pki
riskmandate.ai
risks
sgit.ai
what-can-it-do.games
[Acceptance intervals]
riskmandate.ai
risks
sgit.ai
[Licence to operate]
abp
games
graphs
riskmandate.ai
sgit.ai
store
ungovr.providers
what-can-it-do.games
[Insurability]
abp
riskmandate.ai
sgit.ai
ungovr.providers
what-can-it-do.games
[Security audits]
abp
open-source
riskmandate.ai
sgit.ai
skills
standards
store
teams
threat-modeling
what-can-it-do.gamesStandards and method
mindmap
root((Standards and method))
[EU AI Act]
graphs
newsroom
riskmandate.ai
risks
sgit.ai
standards
ungovr.providers
[GDPR]
graphs
open-source
riskmandate.ai
sgit.ai
standards
teams
ungovr.providers
[Wardley maps]
games
graphs
influences
open-source
pki
riskmandate.ai
sgit.ai
wardley-maps
[Open source]
abp
coding
graphs
issues-fs
llms
newsroom
open-source
riskmandate.ai
risks
sg-compute
sgit.ai
skills
store
twins
ungovr.providersWorking in public
mindmap
root((Working in public))
[Briefs]
abp
llms
riskmandate.ai
sgit.ai
[Interview pages]
riskmandate.ai
sgit.ai
[Lessons learned]
elevenlabs.providers
influences
providers
riskmandate.ai
sgit.ai
teams
ungovr.providers
[Em-dashes and style]
abp
coding
llms
open-source
riskmandate.ai
sgit.ai
[Synthetic users]
riskmandate.ai
sgit.ai
store
[Provenance]
abp
graphs
newsroom
pt.newsroom
sgit.ai
ungovr.providers
[Frozen, hashed sources]
graphs
newsroom
pt.newsroom
riskmandate.ai
sgit.aiThe business
mindmap
root((The business))
[Business plans]
sgit.ai
[Partnerships]
riskmandate.ai
sgit.ai
[Pricing ladders]
elevenlabs.providers
riskmandate.ai
sgit.ai
store
[Startups]
riskmandate.ai
sgit.ai
store
[Brand assets]
riskmandate.ai
ungovr.providersWhat the maps show
- Three concepts are on one site only: vault key management, business plans, and performance and cost are found only on sgit.ai (the concepts page). They are sgit.ai's call, plans and measurements, not yet the network's vocabulary.
- The permission vocabulary travels furthest from where it started. Grants and mandates and blast radius are defined on riskmandate.ai (grant gap↗), undo classes on abp.sgit.ai (undo↗), and between them they reach the two games sites, the provider sites and nhi.sgit.ai.
- Read keys and open source are on the most sites. A read key is how a published vault is opened (credentials↗), so the concept follows the vaults from site to site.
Where a concept is defined is on the concepts page. How the sites link to each other, whatever the concept, is the network map.