Source authority ranking
The machine has to know the difference between binding authority and commentary.
A well-written memo can be more persuasive on the page than the regulation it is describing, and a retrieval system that ranks by similarity alone will hand you the memo. Authority is therefore declared, not inferred from how confident a document sounds.
Tier 1
Binding authority
Statute, regulation, court order, certified standard in force.
0 source(s) here
Tier 2
Official guidance
Agency manual, advisory, published procedure. Persuasive, not binding.
0 source(s) here
Tier 3
Primary record
A dataset, budget, filing or measurement. Reports facts, asserts no rule.
5 source(s) here
Tier 4
Secondary research
Analysis, memo, article. Cites authority; is not authority.
8 source(s) here
Tier 5
Draft or working note
Not adopted. May contradict the finished version and often does.
7 source(s) here
—
Unranked
The catalog was never told. This is not the same as weak.
0 source(s) here
Nothing in this staging folder is binding authority. There is no tier-1 source here, and the studio will not promote one to fill the gap. Everything above is a specification, a model, a measurement or a strategy report — useful, and not law. A jurisdiction database of statutes and regulations is exactly what tier 1 is reserved for, and none is connected yet.
| Tier | Source | Rank | Why it ranks there | Freshness |
|---|---|---|---|---|
| 3 | Editorial taxonomy03_DATASETS/Editorial_Taxonomy.csv | Primary record | The section structure as decided. | FRESH verified today |
| 3 | Product families03_DATASETS/Product_Families.csv | Primary record | A record of a decision taken, not an analysis of one. | FRESH verified today |
| 3 | Risk register03_DATASETS/Risk_Register.csv | Primary record | A register of positions held, with mitigations assigned. | FRESH verified today |
| 3 | Role OS catalog03_DATASETS/Role_OS_Catalog.csv | Primary record | A record of a decision taken. | FRESH verified today |
| 3 | BB2G subdomain estate — live probe, 15 August 202603_DATASETS/bb2g_fleet_probe_2026-08-15.json | Primary record | A measurement, taken by this build, of what actually answered. It reports; it asserts no rule. | FRESH verified today |
| 4 | Intelligence Workshop — package guide00_START_HERE.txt | Secondary research | It describes the package. It asserts no rule and reports no measurement. | FRESH verified today |
| 4 | Deep research appendices04_RESEARCH/Expanded_Deep_Research_Appendices.md | Secondary research | Research notes. They cite authority; they are not authority. | FRESH verified today |
| 4 | Maricopa rollout — cost to deliver and what you can charge04_RESEARCH/Maricopa County Election Rollout — Cost to Deliver and What You Can Charge.md | Secondary research | It says so itself: county spending is sourced to primary documents, everything about her own cost is a labelled model. | FRESH last verified 1 day ago |
| 4 | Research source list04_RESEARCH/Research_Sources.txt | Secondary research | A bibliography. | FRESH verified today |
| 4 | Building the Intelligence Workshop05_BOOKS_AND_DOCUMENTS/Building The Intelligence Workshop_ A Productized Media, Software, Education, and Enterprise Ecosyst.docx | Secondary research | A strategy report with a trademark finding in it. | FRESH verified today |
| 4 | Intelligence Workshop master blueprint (markdown)05_BOOKS_AND_DOCUMENTS/Intelligence_Workshop_Master_Blueprint.md | Secondary research | A strategy report. | FRESH verified today |
| 4 | Election Command and LockChain — production build blueprint06_APP_SPECIFICATIONS/BB2G Election Command and LockChain_ Full Production Build Blueprint.docx | Secondary research | A research-backed build recommendation. | FRESH last verified 11 days ago |
| 4 | Sovereign AI Shield — architecture, security model, compliance, product pack06_APP_SPECIFICATIONS/Sovereign AI Shield_ Deployable Architecture, Security Model, Compliance Framework, and Product Pack.docx | Secondary research | A vendor-cited architecture study. | FRESH verified today |
| 5 | Community channel architecture03_DATASETS/Discord_Architecture.csv | Draft or working note | A plan for a space that does not exist yet. | FRESH verified today |
| 5 | Pricing tiers03_DATASETS/Pricing_Tiers.csv | Draft or working note | It is a proposal. No tier here has been adopted, and one line contradicts the standing access policy. | FRESH verified today |
| 5 | Revenue scenarios03_DATASETS/Revenue_Scenarios.csv | Draft or working note | A model, and two of its three rows do not add up. | FRESH verified today |
| 5 | 90-day roadmap03_DATASETS/Roadmap.csv | Draft or working note | A plan. Phases carry no start date. | FRESH verified today |
| 5 | Mail-ballot lifecycle token — complete specification v1.006_APP_SPECIFICATIONS/complete_mail_ballot_lifecycle_token_specification.md | Draft or working note | Its own status line reads: Draft architecture and implementation specification. | UNVERIFIED this source has never been verified against its origin |
| 5 | Mail-ballot lifecycle token — design06_APP_SPECIFICATIONS/mail_ballot_lifecycle_token_design.md | Draft or working note | Its own status line reads: architecture and implementation starter specification. | UNVERIFIED this source has never been verified against its origin |
| 5 | Mail-ballot lifecycle token — unified specification v1.106_APP_SPECIFICATIONS/unified_mail_ballot_lifecycle_token_specification.md | Draft or working note | Its own version line reads: 1.1 draft. It is the latest, and it is still a draft. | UNVERIFIED this source has never been verified against its origin |
Version and effective-date engine
Never silently mix current and superseded.
Three documents in this staging folder specify the same system. Left in a plain document index they would all come back to the same query, and the answer would be a blend of a starter sketch, a v1.0 draft and a v1.1 draft — with no indication that they were ever different documents. The chain resolves them.
MBLTS — Mail-Ballot Lifecycle Token Service
| Version | Status | Document | Points at |
|---|---|---|---|
| 1.0 | SUPERSEDED | Mail-ballot lifecycle token — complete specification v1.006_APP_SPECIFICATIONS/complete_mail_ballot_lifecycle_token_specification.md | MBLTS-1.1 |
| starter | SUPERSEDED | Mail-ballot lifecycle token — design06_APP_SPECIFICATIONS/mail_ballot_lifecycle_token_design.md | MBLTS-1.0 |
| 1.1 | CURRENT | Mail-ballot lifecycle token — unified specification v1.106_APP_SPECIFICATIONS/unified_mail_ballot_lifecycle_token_specification.md | nothing — it is the head of the chain |
Resolved: MBLTS-starter governs as at 2026-08-15.
Even the head of this chain is a draft. An answer drawn from it names its version, because "the current one" and "an adopted one" are different claims and only the first is true here.
Freshness
For law, regulation, elections, cyber security and government data, stale information is not merely less useful — it is dangerous, because it looks exactly like current information. Every source therefore carries last verified, and where a document carries no date the studio says UNVERIFIED rather than substituting a file modification time, which records when the file was written and nothing at all about whether it is still true.
| Band | Meaning | Sources |
|---|---|---|
| FRESH | Verified within 90 days | 17 |
| AGING | Verified within a year | 0 |
| STALE | Over a year since anyone checked | 0 |
| UNVERIFIED | The document carries no date. Nothing can be said about how current it is. | 3 |