# subscriptions.sgit.ai — everything, in one file Site version: v0.1.0. Generated by admin/build/gen_llms_full.py — do not edit by hand. This is /llms.txt, then the front page, then all ten source documents, concatenated in reading order. It exists because agent fetch tools frequently refuse URLs a search has not already returned, which makes link-following unreliable and makes a single-file surface the practical one. If you can make more than one request, prefer the individual documents at https://subscriptions.sgit.ai/briefs/ — they are the source of truth and this file is a concatenation of them. One document from the source pack is deliberately absent: the source manifest, which indexes referenced corpus material by repository path. All content is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). Third-party material quoted inside these documents stays under its own terms. Contents, in order: 1. llms.txt — the annotated map, each entry carrying its page's single most important fact 2. index.md — the front page as markdown 3-12. the source documents, verbatim ============================================================================== == llms.txt ============================================================================== # subscriptions.sgit.ai — a subscription is a discount, not rent > A subscription is a discount for committing to regular use. It is not rent on > something you have the right to ignore. Site version: v0.1.0 (25 August 2026). All content CC BY 4.0. ## The operating insight The first question is not "how much did I use it?" — which most companies cannot answer and no one can prove. It is "do you hold a record of my usage?" — which every company can answer, is comparable across companies, concedes nothing, and produces a publishable table on day one. ## What kind of site this is Three things, on one domain, by decision of 24 August 2026 — with the commercial service designed to split out later to its own property if a business partner emerges, the riskmandate.ai / risks.sgit.ai pattern run in the other direction: 1. THE STANDARD — five externally-checkable clauses a company can adopt this week, plus the same clauses re-expressed as a proposal for legislation. 2. THE REGISTER — record, never verdict. No adjectives anywhere. Silence is a row, not a blank. "We do not hold that" is recorded neutrally. 3. THE EVIDENCE — exit-path measurements needing no company's cooperation, and the UK legal levers, each dated with a commencement status and a named owner. Consequences an agent can rely on: - No register entry carries a verdict word. The pre-release gate greps every entry against a published blocklist and fails the build otherwise. - No register entry is traceable to a requester. The unit of the register is a company's behaviour, never a person's case. - Every legal claim carries an as-at date and a commencement status. The 2027 regime is not in force yet; a page that omits that is treated as a leak. - The register launches with zero entries. Five items block the first row — solicitor review of the publishing rule and entry one, the selection rule, the refusal interval, the versioned template, and the authority-capture workflow — and all five are published as pending rather than skipped quietly. ## Properties agents may rely on - Every source document is fetchable at a stable constructed URL: /briefs/v0.33.62__subscriptions-brief-pack__.md. This is a promise, not an accident. - The whole site's document set is available in one fetch at /llms-full.txt, for agents whose fetch tools refuse URLs a search has not already returned. - The front page has a markdown twin at /index.md (this file). ## The pages, each with its single most important fact /index.html The thesis, verbatim, at the top of every page. The comparison that makes the register launchable this month: "do you hold a record?" beats "how much did you use it?" on every axis that matters — answerable, short, comparable, concedes nothing, produces a table on day one. /standard/index.html — THE FIVE CLAUSES Each one checkable by somebody other than the company. Not a scoreboard: no scores exist anywhere in the design. The adopters list starts empty, honestly. /standard/proposal/index.html — THE PROPOSAL FOR LEGISLATION The same five clauses re-expressed as obligations, mapped against the spring-2027 regime, explicitly labelled a proposal rather than an enacted text. /register/index.html — THE REGISTER Zero entries at launch. The question, the entry format (no verdict field anywhere), the three answers with no penalty for "we don't hold it", the inconsistency test, and the governance the source instructs be published before entry one: the selection rule, the refusal interval, the template version, and the review process. /exit/index.html — THE EXIT-PATH METHOD Built first because it needs no cooperation: click counts and routes for fifty companies, dated, method published so anyone can re-run it. Not yet run in this release — the method is published, the measurements are not faked. /law/index.html — THE LAW The levers ranked, cancellation-information failures at the top. Dated 25 August 2026, with a named owner and a commencement-status table: the 2027 regime is NOT YET IN FORCE. The card-issuer lever is named as a perimeter question and excluded at launch. /workflows/index.html — WORKFLOWS & VAULTS, THE SGIT SHOWCASE The March 2026 pipeline reused: statements in a vault, an agent working inside it, versioned and shared by token. Three vault stores with three different rules. The honest limit published: vaults buy portability and controlled sharing, not zero knowledge — an agent reading bank statements reads plaintext, which is why bring-your-own-model is the default. /service/index.html — THE SERVICE (LATER) Thin by design. The future-commercial directory that moves at the split. /shipped/index.html What this site asserts versus what it can demonstrate: built, position, described-from-the-corpus, proposed-and-not-built, and pending-on-a-human — the five blocking items named plainly rather than absorbed. /network/index.html The sgit.ai network, and where this site's neighbours sit. /documents/index.html The ten source documents published whole, with the raw markdown as the source of truth. /admin/index.html · /admin/comms.html · /admin/versions.html How the site is built and gated; open work and asks back to the project lead; the release history. ## What this site will not tell you - Any characterisation of a named company. Facts only — the pre-release gate enforces this on every register entry. - Any client's personal or financial data, including the founder's own — the March workflow is the exhibit, its data never appears. - Live hostnames, account identifiers or stack addresses. - The card-issuer lever as an offered service, ahead of advice on the perimeter. - Any legal claim without its as-at date and commencement status. ## The sibling sites https://sgit.ai — the parent project: the vault layer, the CLI, the platform docs https://risks.sgit.ai — registers, scoring, evidence — the vocabulary this register's design borrows https://standards.sgit.ai — instruments as graphs; the grounding vocabulary the law pages point at rather than rebuild https://nfrs.sgit.ai — the dated-page and staleness discipline this site's law pages are the sharpest test case for https://open-source.sgit.ai — the standard-as-adoptable-artefact argument; the CC BY rationale ## Source documents /briefs/v0.33.62__subscriptions-brief-pack__README.md — the pack's reading order /briefs/v0.33.62__subscriptions-brief-pack__00__BRIEF.md — the commission, the naming decision, the build order /briefs/v0.33.62__subscriptions-brief-pack__01__the-standard.md — the five clauses /briefs/v0.33.62__subscriptions-brief-pack__02__the-register.md — the question, the schema, the inconsistency test /briefs/v0.33.62__subscriptions-brief-pack__03__the-workflows-and-vaults.md — the sgit showcase /briefs/v0.33.62__subscriptions-brief-pack__04__the-legal-pages.md — the levers, the dates, the perimeter /briefs/v0.33.62__subscriptions-brief-pack__05__site-architecture-and-the-split.md — page by page /briefs/v0.33.62__subscriptions-brief-pack__06__publishing-rules-and-boundaries.md — the register's constitution /briefs/v0.33.62__subscriptions-brief-pack__07__gaps-and-open-questions.md — published unresolved /briefs/v0.33.62__subscriptions-brief-pack__LICENSE.md — CC BY 4.0 Attribution: Dinis Cruz, with AI co-authorship. Repository: https://github.com/SGit-AI/SGit-AI__Website__Subscriptions ============================================================================== == index.md — the front page ============================================================================== # subscriptions.sgit.ai — a subscription is a discount, not rent > A subscription is a discount for committing to regular use. It is not rent on > something you have the right to ignore. Site version: v0.1.0 (25 August 2026). All content CC BY 4.0. ## The operating insight The first question is not "how much did I use it?" — which most companies cannot answer and no one can prove. It is "do you hold a record of my usage?" — which every company can answer, is comparable across companies, concedes nothing, and produces a publishable table on day one. ## What kind of site this is Three things, on one domain, by decision of 24 August 2026 — with the commercial service designed to split out later to its own property if a business partner emerges, the riskmandate.ai / risks.sgit.ai pattern run in the other direction: 1. THE STANDARD — five externally-checkable clauses a company can adopt this week, plus the same clauses re-expressed as a proposal for legislation. 2. THE REGISTER — record, never verdict. No adjectives anywhere. Silence is a row, not a blank. "We do not hold that" is recorded neutrally. 3. THE EVIDENCE — exit-path measurements needing no company's cooperation, and the UK legal levers, each dated with a commencement status and a named owner. Consequences an agent can rely on: - No register entry carries a verdict word. The pre-release gate greps every entry against a published blocklist and fails the build otherwise. - No register entry is traceable to a requester. The unit of the register is a company's behaviour, never a person's case. - Every legal claim carries an as-at date and a commencement status. The 2027 regime is not in force yet; a page that omits that is treated as a leak. - The register launches with zero entries. Five items block the first row — solicitor review of the publishing rule and entry one, the selection rule, the refusal interval, the versioned template, and the authority-capture workflow — and all five are published as pending rather than skipped quietly. ## Properties agents may rely on - Every source document is fetchable at a stable constructed URL: /briefs/v0.33.62__subscriptions-brief-pack__.md. This is a promise, not an accident. - The whole site's document set is available in one fetch at /llms-full.txt, for agents whose fetch tools refuse URLs a search has not already returned. - The front page has a markdown twin at /index.md (this file). ## The pages, each with its single most important fact /index.html The thesis, verbatim, at the top of every page. The comparison that makes the register launchable this month: "do you hold a record?" beats "how much did you use it?" on every axis that matters — answerable, short, comparable, concedes nothing, produces a table on day one. /standard/index.html — THE FIVE CLAUSES Each one checkable by somebody other than the company. Not a scoreboard: no scores exist anywhere in the design. The adopters list starts empty, honestly. /standard/proposal/index.html — THE PROPOSAL FOR LEGISLATION The same five clauses re-expressed as obligations, mapped against the spring-2027 regime, explicitly labelled a proposal rather than an enacted text. /register/index.html — THE REGISTER Zero entries at launch. The question, the entry format (no verdict field anywhere), the three answers with no penalty for "we don't hold it", the inconsistency test, and the governance the source instructs be published before entry one: the selection rule, the refusal interval, the template version, and the review process. /exit/index.html — THE EXIT-PATH METHOD Built first because it needs no cooperation: click counts and routes for fifty companies, dated, method published so anyone can re-run it. Not yet run in this release — the method is published, the measurements are not faked. /law/index.html — THE LAW The levers ranked, cancellation-information failures at the top. Dated 25 August 2026, with a named owner and a commencement-status table: the 2027 regime is NOT YET IN FORCE. The card-issuer lever is named as a perimeter question and excluded at launch. /workflows/index.html — WORKFLOWS & VAULTS, THE SGIT SHOWCASE The March 2026 pipeline reused: statements in a vault, an agent working inside it, versioned and shared by token. Three vault stores with three different rules. The honest limit published: vaults buy portability and controlled sharing, not zero knowledge — an agent reading bank statements reads plaintext, which is why bring-your-own-model is the default. /service/index.html — THE SERVICE (LATER) Thin by design. The future-commercial directory that moves at the split. /shipped/index.html What this site asserts versus what it can demonstrate: built, position, described-from-the-corpus, proposed-and-not-built, and pending-on-a-human — the five blocking items named plainly rather than absorbed. /network/index.html The sgit.ai network, and where this site's neighbours sit. /documents/index.html The ten source documents published whole, with the raw markdown as the source of truth. /admin/index.html · /admin/comms.html · /admin/versions.html How the site is built and gated; open work and asks back to the project lead; the release history. ## What this site will not tell you - Any characterisation of a named company. Facts only — the pre-release gate enforces this on every register entry. - Any client's personal or financial data, including the founder's own — the March workflow is the exhibit, its data never appears. - Live hostnames, account identifiers or stack addresses. - The card-issuer lever as an offered service, ahead of advice on the perimeter. - Any legal claim without its as-at date and commencement status. ## The sibling sites https://sgit.ai — the parent project: the vault layer, the CLI, the platform docs https://risks.sgit.ai — registers, scoring, evidence — the vocabulary this register's design borrows https://standards.sgit.ai — instruments as graphs; the grounding vocabulary the law pages point at rather than rebuild https://nfrs.sgit.ai — the dated-page and staleness discipline this site's law pages are the sharpest test case for https://open-source.sgit.ai — the standard-as-adoptable-artefact argument; the CC BY rationale ## Source documents /briefs/v0.33.62__subscriptions-brief-pack__README.md — the pack's reading order /briefs/v0.33.62__subscriptions-brief-pack__00__BRIEF.md — the commission, the naming decision, the build order /briefs/v0.33.62__subscriptions-brief-pack__01__the-standard.md — the five clauses /briefs/v0.33.62__subscriptions-brief-pack__02__the-register.md — the question, the schema, the inconsistency test /briefs/v0.33.62__subscriptions-brief-pack__03__the-workflows-and-vaults.md — the sgit showcase /briefs/v0.33.62__subscriptions-brief-pack__04__the-legal-pages.md — the levers, the dates, the perimeter /briefs/v0.33.62__subscriptions-brief-pack__05__site-architecture-and-the-split.md — page by page /briefs/v0.33.62__subscriptions-brief-pack__06__publishing-rules-and-boundaries.md — the register's constitution /briefs/v0.33.62__subscriptions-brief-pack__07__gaps-and-open-questions.md — published unresolved /briefs/v0.33.62__subscriptions-brief-pack__LICENSE.md — CC BY 4.0 Attribution: Dinis Cruz, with AI co-authorship. Repository: https://github.com/SGit-AI/SGit-AI__Website__Subscriptions ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__README.md ============================================================================== # subscriptions.sgit.ai — brief pack **For:** the agent commissioned to build `subscriptions.sgit.ai` **From:** Dinis Cruz, via the SG/Send Librarian **Version:** v0.33.62 · 24 August 2026 **Licence:** CC BY 4.0 — but read `LICENSE.md`: this site's real rules are defamation and personal-data discipline, not copyright. --- ## What this is The subscription standard, register and workflows — everything on one site, by decision of 24 August, with the commercial service designed to refactor out later to its own property if a business partner emerges (**the riskmandate pattern, run in the other direction**). Two source briefs ship in `sources/`. The second (*the refund is the pretext and the register is the product*) **overrides the first's framing and keeps all its research**: > *"The law is not the product. It is evidence, and a floor. The product is the record of who answers, what they hold, and how they behave when asked."* And the front-page principle: > **A subscription is a discount for committing to regular use. It is not rent on something you have the right to ignore.** The site also carries the legislative ambition: the five clauses are **the recommended subscription model — the proposal for what the law should be** — published as an adoptable standard first, with the law-shaped draft as the artefact for consumer groups and committees. --- ## Read in this order | File | Words | What it does | |---|---:|---| | **`00__BRIEF.md`** | 1.1k | **Start here.** The two briefs and the override, the decision, the three load-bearing rules, the split table, build order | | **`01__the-standard.md`** | 0.7k | The principle, the five clauses, standard-before-law sequencing, the four-axis maturity model | | **`02__the-register.md`** | 0.6k | Do-you-hold-it, the no-verdict entry format, the three answers, **the inconsistency test**, the reflexivity defences, governance | | **`03__the-workflows-and-vaults.md`** | 0.9k | The March pipeline, the three-vault architecture, the honest vault claim, BYOM-first — **and the sgit showcase argument** | | `04__the-legal-pages.md` | 0.7k | The levers ranked, the closing window, the dating discipline, the perimeter | | `05__site-architecture-and-the-split.md` | 0.5k | Page by page, and the three mechanisms that make the future split a DNS change | | `06__publishing-rules-and-boundaries.md` | 0.5k | The register's constitution as build requirements | | `07__gaps-and-open-questions.md` | 0.6k | 5 blocking items, 9 open questions (3 new), the honest gaps | | `09__source-manifest.csv` | 19 rows | Sources, tiered; corpus paths verified; two marked verify-at-build | | `sources/` | 2 briefs | The 16th and 17th of 20 August, verbatim | --- ## The five things that make this site unusual in the network **1. Its rules bound its own name.** No-adjectives killed every "fair/honest" candidate; the corrected headline killed every "refund" name. The same filters govern every future headline — and the DNR trap (Discount-Not-Rent) is noted so nobody rediscovers it. **2. It can launch before contacting anyone.** The exit-path measurement — fifty companies, click counts, dated, method published — needs **no cooperation**: *"a public, dated, reproducible artefact before a single letter has been sent."* Build it first; it is the register's credibility deposit. **3. Its strongest instrument is the fairest.** The inconsistency test — *"the answer to a data subject and the claims in the marketing cannot both be true"* — lands only where true, needs no characterisation, and passes clean companies cleanly. **4. Its legal pages expire on a known date.** Spring 2027. Every page dated, statused, and owned — the memory-network staleness rule at its sharpest, because here a stale page gets a reader corrected by a trader. **5. It is the sgit business-model showcase.** Client files in client vaults, parsers as shared vault workspaces, the register as re-pointable vault data, BYOM because the meter costs the privacy claim, serverless economics — *"no new infrastructure was invented for this business."* Feature `/workflows/` from `sgit.ai`. --- ## Before anything publishes Five blocking items (`07__` §1): solicitor review of the publishing rule and entry one · the selection rule published · the refusal interval published · the template versioned · the authority-capture workflow. **The register's governance page goes live before its first row.** --- This file is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__00__BRIEF.md ============================================================================== # 00 — The Brief: `subscriptions.sgit.ai` **Version** v0.33.62 · 24 August 2026 **From** Dinis Cruz, via the SG/Send Librarian **To** the agent commissioned to build `subscriptions.sgit.ai` **Licence** CC BY 4.0 — with unusual publishing rules; read `06__` before the first register entry --- ## 1. The commission, and the decision that shapes it Two source briefs ship in `sources/`, and **the second overrides the first's framing while keeping all of its research**: - **The 16th of 20 August** — the claims service: recover money from subscription compliance failures. Its legal mapping is accurate and survives in full. - **The 17th of 20 August** — the redirect: *"the law is not the product. It is evidence, and a floor. **The product is the record of who answers, what they hold, and how they behave when asked.**"* The naming decision, taken 24 August: **everything on `subscriptions.sgit.ai`** — the standard, the workflows, the register results — with the commercial service refactored out later to its own property **if a business partner emerges**, exactly the `riskmandate.ai` / `risks.sgit.ai` pattern. `05__` designs that split in from day one, so it is a DNS change rather than a migration. **And the site carries a legislative ambition:** the five-clause standard is *"our recommended subscription model — basically our proposal for legislation."* Draft as a standard first (*"a standard can be adopted by one company next week and a law cannot be adopted by anybody"*), with the law-shaped version as the artefact for consumer groups and committees. --- ## 2. The thesis Front page, top, verbatim: > **A subscription is a discount for committing to regular use. It is not rent on something you have the right to ignore.** And the operating insight that makes the site launchable this month: > The first question is not *"how much did I use it?"* — which most companies cannot answer and no one can prove. It is ***"do you hold a record of my usage?"*** — which **every company can answer**, is comparable across companies, concedes nothing, and produces a publishable table on day one. --- ## 3. The three load-bearing rules These are the site's constitution. Every page obeys them; `06__` enforces them. **(a) Publish the record, never the verdict.** *"No adjectives anywhere."* Two independent arguments converge on it: the Defamation Act 2013 (serious-financial-loss threshold; substantial truth a complete defence — a dated record is protected, a characterisation is exposed) and the corpus's own discipline (*"a dated test somebody can repeat is evidence, and an assertion from a participant is marketing"*). *"A page that says asked on 3 March, replied on 2 April, said no, is more damaging than any adjective and it is safe."* **Note: this rule bound the site's own name** — no "fair", no "honest" — and it binds every headline. **(b) "We do not hold that" is a legitimate answer, recorded neutrally.** Data minimisation makes not retaining usage logs *correct* behaviour; scoring it as evasion would punish privacy and discredit the register in one exchange. The power move is the **inconsistency test**: *"the answer to a data subject and the claims in the marketing cannot both be true"* — a company claiming personalisation and engagement analytics while holding no usage record fails on facts, not adjectives. **(c) Every access request is genuine, never leverage.** The regulator names leverage as the one refusable case, so merit is tested **before** the request, never after — and the service publishes its decline rate: *"a service that pursues everything is not assessing anything."* --- ## 4. What the site holds — and which half moves later | Content | CC BY, stays forever | Moves at the split | |---|:-:|:-:| | **The standard** — five clauses, each externally checkable · the law-shaped draft | ✅ | | | **The legal pages** — the levers table, dated, with commencement status | ✅ | | | **The workflows** — statements → inventory → evidence → claim, agentic, plaintext step stated | ✅ | | | **The register** — record-only entries, the schema, the templates | ✅ *(data in a vault → re-pointable to a neutral domain / non-profit)* | | | **The exit-path measurements** — method + results | ✅ | | | **The parsers** — per-provider workspaces | ✅ | | | The recovery *service* — intake, fees, client vaults, hosted metering | | ✅ | | Pricing, the percentage model, the decline-rate dashboard | | ✅ | --- ## 5. The honesty constraints — all from the sources' own words - **"You did not use it" yields nothing on its own** — and *that row goes on the website, in those words*, because *"a service that says plainly what it cannot do is the only kind anybody should trust with their bank statements."* - **The 2027 regime is not in force.** Every legal page carries an as-at date and commencement status; *"a page describing cooling-off rights that does not say 'not in force until spring 2027' is actively harmful."* - **Vaults buy portability and controlled sharing, not zero knowledge.** *"An agent that reads bank statements to find subscriptions is reading plaintext."* Publish that row. - **The card-issuer lever may sit inside the regulated claims perimeter** — launch without it, add after advice. - **No solicitor has reviewed any of this** — and the publishing rule must be reviewed **before** the first entry, not after. - **The founder is the wrong sample** — technical, persistent, high subscription count; *"the second and third clients should be chosen for being unlike him."* --- ## 6. The numbers | | | |---|---| | **Sources** | 2 briefs, 20 Aug 2026 (~9,000 words), + 6 corpus briefs referenced and verified on disk | | **The standard** | 5 clauses, all externally checkable · 3 answer types · 4 maturity axes, 1 needing no cooperation | | **The window** | 12-month reach-back live now · regime lands **spring 2027** (slipped from autumn 2026) · UCP enforcement in force since 6 Apr 2025 | | **The pipeline** | Built March 2026 on the founder's own statements — *"only the destination changes"* | | **This pack** | 8 documents + both source briefs · manifest of 18 rows | --- ## 7. Build order — from the source's own "what ships first" 1. **The principle**, published, five clauses, each with how it is checked. 2. **The exit-path measurements for fifty companies** — *"no company needs to cooperate and it can be done this week"*, and it is the argument for the whole project: a public, dated, reproducible artefact before a single letter is sent. 3. **The question**, as one published template, versioned. 4. **The register schema** — three answers distinguished, no verdict field anywhere. 5. **Ten companies asked by the founder**, table published including the silences. 6. **The parsers**, as contributors add them. The acceptance test, verbatim: *"A stranger can read the site, adopt the principle, run the exit-path measurement on a company nobody has covered, send the question using the published template, and add a row to the register, without asking anybody for permission or explanation."* --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__01__the-standard.md ============================================================================== # 01 — The standard: five clauses, and the proposal for legislation The site's anchor page, and the reason everything lives here rather than on a service brand: **this is the recommended subscription model, published as a standard a company can adopt this week, with the law-shaped version as a later artefact for consumer groups and committees.** --- ## 1. The principle > **A subscription is a discount for committing to regular use. It is not rent on something you have the right to ignore.** *"That sentence does the work of the whole document, and it is the one to put at the top of the site."* It is the tagline of every page — and per the naming discussion, it stays a tagline rather than a name (the abbreviation trap is real: Discount-Not-Rent → DNR). ## 2. The five clauses Each one a company either meets or does not, **each observable from outside** — which is *"the property that makes this a standard rather than a pledge."* | # | Clause | Observable how | |---|---|---| | 1 | **We can tell you how much you used it** | Ask, and see | | 2 | **We tell you without you having to ask** | Do usage summaries arrive? | | 3 | **Leaving is as easy as joining, by the same route** | Testable in a browser, without their cooperation | | 4 | **If you stop using it, we tell you before we charge you again** | Observable over one renewal cycle | | 5 | **We do not charge for a period you could not have used** | Outages, suspensions, lockouts | The corpus discipline behind the design: *"a dated test somebody else can repeat is evidence, and an assertion is marketing. A code of conduct whose clauses cannot be checked is the second thing."* ## 3. The standard-before-law sequencing *"Draft it as a standard rather than as legislation, because a standard can be adopted by one company next week and a law cannot be adopted by anybody."* So the page ships in three layers: 1. **The standard** — the five clauses, adoptable unilaterally, with a public adopters list (empty at launch, honestly). 2. **The conformance route** — how a company declares adoption: **a published policy page**, which the register accepts as an answer (the reflexivity fix: answering becomes a one-time cost). 3. **The draft law** — the same five clauses re-expressed as obligations, written later, explicitly labelled as a proposal, with the spring-2027 regime mapped beside it showing what it covers and what the five clauses add. Layer 3 is where the site's ambition lives, and it has an unusually strong position: by the time the 2027 regime lands, the site will hold a year of register data showing which clauses companies already meet voluntarily — **evidence for the legislative proposal generated by the register itself.** ## 4. The maturity model — four axes, one needing no cooperation | Axis | The question | Observed how | |---|---|---| | **Knows** | Can they tell you what you used? | Ask | | **Tells** | Do they tell you before charging you again? | Watch one renewal cycle | | **Speed** | Days from asking to a substantive answer | Measured | | **Exit** | Is leaving as easy as joining, by the same route? | **A browser, no contact required** | **Exit is built first** — click counts and routes for fifty companies, dated, method published so anyone can re-run it. *"It is measured rather than reported, so no company can decline to participate in it."* It follows the reproducible-test discipline of the 16 August comparison-pages brief, and it is the register's credibility deposit: a complete artefact before a single letter is sent. ## 5. What the standard is not Not a scoreboard (no scores exist anywhere in the design) · not a pledge (every clause is externally checkable) · not the 2027 regime restated (three of the five clauses go beyond it — that delta is the legislative argument) · and not dependent on anyone's cooperation to start (clause 3 and the exit axis are measurable today, unilaterally). --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__02__the-register.md ============================================================================== # 02 — The register The product. *"The refund is the pretext that generates the data and the record of who answers is the asset."* --- ## 1. The question, and why it is the design's sharpest move > **"Do you hold a record of my usage of this service?"** Non-use is unprovable in the direction that matters; whether they *hold* the record is answerable by every company, short, comparable, concedes nothing, and produces a table on day one. The second question — "then show me" — follows only where the first is yes, *"because a company that has said it holds the record has no basis for withholding it from the person it is about."* ## 2. The entry format — no verdict field anywhere | Field | Note | |---|---| | Company | As named on the contract | | Date asked | | | Exactly what was asked | Quoting the published template, **by version** | | Date replied, or no reply as at a date | **Silence is a row, not a blank** | | What was said | **Verbatim, requester's details removed** | | Nothing else | **No score, no label, no adverb** | The legal footing, checked in the source: serious-financial-loss threshold for trading bodies; substantial truth a complete defence; honest opinion only where the basis is published beside it. *"The record is that basis"* — and the design avoids the invitation rather than relying on winning, because *"the deterrent in practice is the cost of defending rather than the likelihood of losing."* ## 3. The three answers — three different things | Answer | What it is | Recorded | |---|---|---| | We hold it, here it is | The best case | With the response time | | We hold it and will not give it | A compliance question about a right the person has | Verbatim, with the reason given | | **We do not hold it** | **A legitimate design choice** (data minimisation) | **Neutrally, no penalty** | And the third row's sting, which is the register's strongest instrument and its fairest: **the inconsistency test.** *"A company that holds no usage record cannot simultaneously be personalising your experience, recommending content based on what you watched, optimising engagement, or reporting engagement metrics to investors… The test is not that they refused. It is that the answer to a data subject and the claims in the marketing cannot both be true."* Factual, checkable by anybody, lands only where true — *"a company that genuinely holds nothing and claims nothing passes cleanly."* ## 4. The reflexivity defences — designed in, not discovered A register that publishes refusals teaches legal teams to refuse. Three mitigations, in order of leverage: 1. **A published policy page counts as an answer** — answering becomes a one-time cost, *"the only version a large company will actually do."* 2. **The good path published as loudly as the bad** — *"answered in full, in nine days, in machine-readable format"* as prominent as any refusal; *"a site that only names other people's gaps is not read as research."* 3. **Never publish a first refusal** — ask twice, stated interval, so *"every published refusal is a considered one."* ## 5. Governance — the criticisms, pre-empted by publishing the rules The source names the criticism that arrives first: *"a founder-led list is a founder's grievances until the selection rule is published."* So the register publishes, before entry one: the **selection rule** for who gets asked · the **interval** defining a refusal · the **template versions** · the **review process** for the first entry (a solicitor, before, not after) · and the **transition intent** — the register data lives in a vault, so the move to a neutral domain or non-profit is a re-point, not a migration. --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__03__the-workflows-and-vaults.md ============================================================================== # 03 — The workflows, and why this is the sgit showcase The commission's third note: *"this is also a good example of the kind of companies and business models that can be built on top of sgit and encrypted vaults + Vault Apps."* This document is that argument, made concretely — and the workflows page is what makes the site *"more than a law summary."* --- ## 1. The pipeline exists — March 2026, the founder's own statements The 29 March debrief records the exact workflow, already run: credit card statements shared into a vault by short token, an agent cloning it, **working inside the vault**, committing and pushing back, the human pulling and seeing exactly what changed — *"a genuine shared workspace, persistent and versioned across sessions."* | The March workflow | This service | |---|---| | Statements as raw input | The same | | Parse into a ledger | Parse into a **subscription inventory** | | Agent works inside the vault, versioned | Unchanged | | Shared by a short token | Unchanged | | Output: an accountant-ready pack | **Output: a claim file per subscription** | *"Only the destination changes… the novel work is the legal mapping and the claim file, not the plumbing."* And the April positioning brief already named the audience — people who send document bundles to accountants and brokers, *"folders that remember everything"* — and the economics: *"a serverless, low-cost business."* ## 2. The vault architecture — three stores, three rules | Store | Holds | Rule | |---|---|---| | **The person's own vault** | Statements, the claim file, full responses, anything identifying them | *"A completed claim file is a concentrated dossier of one person's financial life"* — it belongs to them; the operator holds *"only what a live claim requires"* | | **The public register vault** | Company rows, dates, verbatim replies (details removed), exit measurements, templates, parsers | **"The unit of the register is a company's behaviour, never a person's case"** — CC BY, re-pointable to a neutral domain at the split | | **Per-provider parser workspaces** | The template for that provider, the known reply shape, the extraction workflow | *"One person works out how to parse one provider's export, and everybody after them gets it for nothing"* — **the crowdsourced asset is the parsers and the templates, not the data** | ## 3. The honest vault claim — publish the last row | Property | True? | |---|---| | The client holds their own evidence pack and can take it elsewhere | **Yes** | | Sharing with a specific party is a key rather than an account | **Yes** | | The store cannot read it | **Yes** | | Every change versioned — what was submitted when is answerable **in a dispute** | **Yes** | | The agent reading the statements cannot see the contents | **No — it reads plaintext** | *"Vaults buy portability and controlled sharing, not zero knowledge… the privacy claim for the processing step is operational (we do not retain it) rather than architectural (we cannot see it)."* Publishing that row is the estate's standing discipline — *"a page that names where its own approach loses is what makes the rest of it credible."* **Which is why bring-your-own-model is the default**, inverted from the usual ordering: *"being the meter puts you in the request path"*, and the content here is bank statements. BYOM keeps the claim architectural; hosted-metered is the convenience option, priced per extraction, retention stated structurally. The consumption ledger already exists, generic on unit type by design. ## 4. The SAR workflow, with its constitution The access request is the discovery engine — representative requests must be honoured, high refusal threshold, one month — with the ask-list ordered by claim value: **usage and access logs first** (converts a feeling into a record), then the sign-up record, consent records, billing history, correspondence, and **the terms in force at the time** (not today's website). The one rule that keeps the mechanism alive, verbatim: *"Every access request is a genuine request for the person's own data, sent because they want it. It is never sent as leverage, never bundled with a demand, and never sent in a volume calculated to burden."* Merit is tested **before** the request; the decline rate is published. Authority to act is captured as evidence, not assumed. And the portability right is noted for volume: where it works it yields machine-readable output. ## 5. The showcase summary — for `sgit.ai` and the memory network One paragraph the estate can reuse: a consumer service where **the client's file lives in the client's vault**, the operator is thin and serverless, the **shared asset is workflows in vaults** (parsers, templates), the register is **published data in a re-pointable vault**, metering is a generic ledger, the model is **BYOM-first because the meter costs the privacy claim** — and the whole thing was assembled from primitives that already existed: the March pipeline, the token gateway finding, the comparison-page discipline, the store-the-choices rule. **No new infrastructure was invented for this business.** That is the sgit business-model argument in one worked example. --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__04__the-legal-pages.md ============================================================================== # 04 — The legal pages: the levers, the dates, the perimeter The first brief's research, given its correct role: **evidence and a floor.** These pages make the site useful to a claimant today — and they are the pages with a known expiry date, which shapes how they must be published. --- ## 1. The levers, ranked — publish the table whole, including the last row | Lever | Yields | Evidence needed | |---|---|---| | **Cancellation information not given** | Cancellation up to **12 months** beyond the normal period, **and nothing owed for services supplied in it** | The sign-up flow *as it was*, the confirmation, the terms served at the time | | Payments taken after cancellation | Full return | The instruction, its date, the payments after | | An unfair term | The term does not bind | The terms accepted, and how presented | | Care-and-skill failure | Repeat performance or price reduction | Records of what went wrong, when | | Misleading action/omission (UCP, in force Apr 2025) | Enforcement and complaint lever | The flow, screenshots, the omission | | Card issuer claim | Against the lender too | ⚠️ **perimeter question — launch without it** | | Chargeback / DD indemnity | Reversal / bank return | Statements, and speed | | **"You did not use it"** | **Nothing on its own** | Useful as evidence, never as the claim | *"The top row is the product"* — it ends the contract *and* returns money, it is *"arithmetic once the evidence exists"*, and the evidence is a screenshot rather than an opinion. *"And the last row is the one to put on the website, in those words."* ## 2. The closing window — the honest version of the pitch The claim pool is **largest now**: traders have not yet been forced to tidy sign-up flows, the reach-back is ~13 months, competitors are few. It shrinks from spring 2027 — and the source names the discomfort squarely: *"the product's best year is also its least defensible pitch."* Publish the window as a fact with its tension, not as urgency marketing. The second act is real: the method transfers to a larger surface with clearer obligations, arriving *"with a year of worked cases while everybody else is reading the guidance."* ## 3. The dating discipline — this domain's special obligation Every legal page carries: **an as-at date · a commencement status · the source links · a named owner**. The status table itself is content: | | Status | |---|---| | Cooling-off on trial conversion / auto-renewal · renewal reminders · mandatory easy exit | **Not in force. Spring 2027** (slipped from autumn 2026 — written statement, 13 Apr 2026) | | Unfair commercial practices + direct enforcement (up to 10% global turnover) | **In force since 6 Apr 2025** | | Distance selling cancellation regime | **In force since 2014 — the workhorse** | *"A page describing cooling-off rights that does not say 'not in force until spring 2027' is actively harmful"* — and the open question from the source stands: *"who owns the dated legal pages? They expire on a known date and a stale page is worse than none."* This is the memory-network staleness rule at its sharpest: a stale legal page is a false memory that gets a reader corrected by a trader. ## 4. The perimeter General consumer refunds sit **outside** regulated claims management (the six sectors checked). **The card-issuer lever is the one row that may cross** — it looks like a financial-services claim — so: launch without it, take advice, add later. Two notes carried into that conversation: contingency fees change how the service looks to a regulator regardless of sector, and representative authority must be *captured as evidence*, not assumed. ## 5. What is not retracted, and what is The first brief's framing (law-as-product) is superseded; **none of its research is.** The site records that correction openly — *"the corpus's value depends on this being said rather than absorbed"* — and the correction itself is a good page: it is the estate's brief-supersedes-brief discipline, demonstrated on the site's own founding documents. --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__05__site-architecture-and-the-split.md ============================================================================== # 05 — Site architecture, and the split designed in from day one ## The decision this implements Everything on `subscriptions.sgit.ai` now; the commercial service refactors out later to its own property **if a business partner emerges** — the `riskmandate.ai` / `risks.sgit.ai` pattern, run in the other direction (they split content out of a commercial site; this starts unified and splits the commercial part off). **Design for the split now and it is a DNS change; ignore it and it is a migration.** Three mechanisms make the split cheap: 1. **The register data lives in a vault** — the site renders it. Re-pointing the render target moves the register to a neutral domain or non-profit without touching an entry. This also answers the founder-led-register criticism structurally: the data was never coupled to the founder's domain. 2. **The service pages live under one path** — `/service/` — and nowhere else links into their internals. The split is an export of one directory plus a stub. 3. **The licence boundary is drawn now** — everything outside `/service/` is CC BY forever; `/service/` content is CC BY too while it lives here, but its *branding and pricing* are the parts a partner would own, so keep them thin. ## Page by page **`/`** — the principle as the epigraph, the do-you-hold-it question, the current register table (even at ten rows), and the last-row honesty: *"you did not use it" yields nothing on its own.* **`/standard/`** — `01__`. The five clauses, the adopters list, the conformance route (policy page = answer), and **`/standard/proposal/`** — the law-shaped draft, explicitly labelled a proposal, with the 2027-regime delta table. This is the legislative artefact. **`/register/`** — `02__`. The table, the schema, the templates by version, the governance page (selection rule, interval, review process) published **before entry one**. **`/exit/`** — the exit-path measurements: method first, then fifty companies, dated, re-runnable. **The first thing built.** **`/workflows/`** — `03__`. The agentic pipeline with the plaintext step stated, the vault architecture, the parser workspaces, BYOM-first. This page doubles as the sgit showcase and should be linked from `sgit.ai` and the `services.sgit.ai` gallery when that exists. **`/law/`** — `04__`. The levers table, the status table, as-at dates, named owner. **The pages with an expiry date.** **`/service/`** — the future-commercial directory: intake, the merit test, the decline rate, fees. Thin at launch; the founder's ten cases run through it. **`/shipped/`** — what exists vs designed, per house pattern: at launch, the standard and the method exist; the register has N rows; the parsers have M providers; the service has one client, who built it. **`/network/`, `/admin/`** — house pattern; `/llms.txt` self-sufficient; `/llms-full.txt`; markdown twins. ## Boundaries within the network | Site | Relationship | |---|---| | `standards.sgit.ai` | Owns instruments-as-graphs; **this site's law pages link its method** (dated provisions, "derived not canonical") rather than rebuilding it | | `risks.sgit.ai` | The acceptance/evidence vocabulary; the claim file is a small evidence graph | | `nfrs.sgit.ai` | The dated-page/staleness discipline — this site is its sharpest test case | | `open-source.sgit.ai` | The standard-as-adoptable-artefact argument; CC BY rationale | | `llms.sgit.ai` / `coding.sgit.ai` | The agentic workflow conventions the parsers follow | | `sgit.ai` | **Should feature `/workflows/` as the business-model worked example** | --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__06__publishing-rules-and-boundaries.md ============================================================================== # 06 — Publishing rules and boundaries This site has the strictest publishing rules in the network, because it publishes **other companies' words** and handles **individuals' financial lives**. The rules are the product's legal armour; treat them as build requirements. ## 1. The register's constitution — enforced, not remembered 1. **Record, never verdict. No adjectives anywhere** — in entries, headlines, URLs, or metadata. A CI check should grep register pages for a published adjective blocklist; cheap, and it makes the rule structural. 2. **Verbatim replies, requester details removed.** The redaction is a pipeline step, not an editorial pass. 3. **Silence is a row, not a blank.** "No reply as at [date]" is a fact. 4. **"We do not hold that" is recorded neutrally.** The inconsistency test is run only against the company's own published claims, quoted and linked. 5. **Never publish a first refusal.** Two asks, published interval. 6. **A published policy page counts as an answer** and is recorded as such. 7. **A solicitor reviews the publishing rule and the first entry before it goes live.** The source flags this explicitly; it is the one external dependency in the launch list. ## 2. Personal data — the hard lines - **The unit of the register is a company's behaviour, never a person's case.** No register entry is traceable to a requester. - **Claim files live in the client's own vault.** The operator holds *"only what a live claim requires"*. Nothing personal in browser storage; nothing personal in the repo; nothing personal in the register vault. - **Real-world examples are published with no personal data** — the source's instruction verbatim. - **SARs are never leverage** — merit before request, decline rate published, authority evidenced. ## 3. Licensing Site content **CC BY 4.0**, stamped and gated as across the network. The register data: CC BY, in its own vault (the re-pointable asset). Company replies quoted in entries are **quotation for reporting** — verbatim, attributed, dated; that is the substantial-truth posture and also fair-dealing hygiene. The two source briefs ship in `sources/` under their own CC BY notices. ## 4. Do not publish - **Any client's personal or financial data** — including the founder's own statements (the March pipeline ran on them; the *workflow* is publishable, the statements never). - **Any characterisation of a named company** — the constitution, restated as a redaction rule. - **The card-issuer lever as an offered service** — perimeter advice first. - **Legal pages without their as-at date and status** — an undated legal page is treated as a leak, not a draft. - **Unreviewed first entries** — rule 7 above. - From the wider estate: the standard exclusions (alchemist tree, named-individual GRC records) apply to any corpus material quoted. ## 5. The tensions, published Per house pattern, the site publishes its own tensions — the source supplies them: the refund-as-pretext (customers want the thing that is not the product) · record-only will feel toothless to some readers · neutral data-minimisation treatment removes the easy headline · the inconsistency test is slow because it requires reading marketing carefully · the good-path-cheap rule lets a company answer with one page and no change · founder-led selection until the rule is published · **and the closing window: the best year is the least defensible pitch.** --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__07__gaps-and-open-questions.md ============================================================================== # 07 — Gaps and open questions ## 1. Before launch — blocking | # | Item | Why | |---|---|---| | **B1** | **Solicitor review of the publishing rule and entry one** | The source's own flag: *"review costs less before than after"* | | **B2** | **The selection rule, published** | *"A founder-led list is a founder's grievances until the selection rule is published"* | | **B3** | **The refusal interval, published** | It defines a refusal; per-case choice is indefensible | | **B4** | **The template, versioned** | Entries quote it by version; it must exist before ask one | | **B5** | **Authority-capture workflow** | Representative SARs need evidence of authority; *"doing it badly is how the mechanism gets refused"* | ## 2. Open questions — carried from the sources plus three new | # | Question | Status | |---|---|---| | **Q1** | Does a published policy page count as an answer even when it means a company answers *"without ever knowing anything"*? | The source's own sharpest tension on the reflexivity fix. Recommendation: yes, recorded as "answered by policy" — a distinguishable row type | | **Q2** | Card-issuer lever: in or out of the perimeter? | Advice, then decide. Launch without | | **Q3** | How is the sign-up flow evidenced retrospectively? | Archives are the only source; a `/method/` page on flow archiving is worth writing early — it is also what makes the exit measurements re-runnable | | **Q4** | What happens at spring 2027? | Planned, not absorbed: the standard's delta table (`01__` §3) is the transition artefact | | **Q5** | Who owns the dated legal pages? | Named owner per page, or the pages do not ship | | **Q6** | When does the register move to a non-profit? | *"Easier before there is a register than after"* — the vault design makes it a re-point; the decision is still a decision | | **Q7** *(new)* | **Jurisdiction signalling** — the levers are UK; the standard and the exit measurements travel. Does the site present as UK-first with a general standard, or general-first with UK levers? | Unanswered from the naming conversation. Recommendation: general standard, clearly-badged UK law section — the register schema is jurisdiction-free | | **Q8** *(new)* | Does the founder's own ten-company ask list pass the selection rule he publishes? | The first test of B2, and worth stating that it was applied | | **Q9** *(new)* | Where do *international* clause-3 measurements go? | Exit paths are measurable anywhere; the register question is UK-flavoured. Decide whether `/exit/` is global from day one | ## 3. Honest gaps - **No adopters at launch.** The standard's adopters list starts empty; publish it empty rather than seeding it with sympathetic friends. - **The service half is one client deep**, and he built it. The unrepresentative-sample warning is in the source; the second and third clients are chosen for being unlike him. - **The inconsistency test needs a method page** — reading marketing *"carefully enough to be sure"* is slow and should be documented as a repeatable procedure before it is run in anger. - **The exit measurements age fastest** — *"a flow can change overnight"* — so the re-run cadence is part of the method, not an afterthought. --- This document is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). ============================================================================== == briefs/v0.33.62__subscriptions-brief-pack__LICENSE.md ============================================================================== # Licence ## This pack Everything in this brief pack — the eight numbered documents, `09__source-manifest.csv`, the two source briefs in `sources/`, this file and `README.md` — is released under the **Creative Commons Attribution 4.0 International licence (CC BY 4.0)**. Copyright (c) 2026 Dinis Cruz Licensed under CC BY 4.0 — https://creativecommons.org/licenses/by/4.0/ Attribution: **Dinis Cruz**, with AI co-authorship (Claude, Anthropic). Both source briefs carry their own CC BY notices. ## The site this pack commissions **The entire content of `subscriptions.sgit.ai`** is CC BY 4.0, stamped and gated with `licence-audit.py --check`. **The register data lives in its own vault under CC BY** — that is the re-pointable asset for the future neutral-domain or non-profit move. ## The rules that outrank the licence This site's exposure is not copyright; it is defamation and personal data. Three rules are build requirements, not editorial preferences: 1. **Publish the record, never the verdict — no adjectives anywhere**, enforced by a CI blocklist on register pages. Company replies are quoted **verbatim, attributed, dated, with requester details removed** — the substantial-truth posture. 2. **No personal data, ever**: no client's statements or claim contents (including the founder's own — the March *workflow* is the exhibit, its data never appears), nothing in browser storage, no register entry traceable to a requester. 3. **A solicitor reviews the publishing rule and the first entry before it goes live.** The one external dependency in the launch list, flagged by the source itself. ## Do not publish The card-issuer lever as an offered service (perimeter advice first) · any legal page without its as-at date and commencement status · any first refusal (two asks, published interval) · the payments-platform pricing detail and the alchemist tree (standard exclusions). ## Accuracy **The legal pages expire on a known date.** Spring 2027 changes the surface; every claim carries its date and source. Two referenced corpus briefs (16 Aug comparison-pages, 20 Aug user-section) post-date this clone — verify their paths at build time, as marked in the manifest. --- This file is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0).