How Vigil Professional works

The principle

Public Vigil is for understanding the system. Professional is for working with it. Professional consumes the same canonical corpus — the same questions, observables, ledgers, events, records and sources — and adds tools, depth and new authored material on top. It restates no assessment, carries no separate truth, and cannot be more confident about anything than the public record is. Where it adds content, that content is a new canonical record with its own review discipline, held to the standard in §10 of the methodology.

Nothing public moves behind the gate. Read, Explore, Map, Watch, the public Tempo series, Notes, Briefs, Profiles, Atlas and the public Standing Watchdemonstration remain public and are not reduced to make room for Professional.

The registers

Every statement a subscriber meets carries one of these labels, printed beside it. A label is never conveyed by colour alone, and no surface may present one register as another.

FACT

A dated, sourced occurrence the corpus records. It says what happened, not what it means.

ASSESSMENT

Vigil’s graded reading of the evidence, carrying its own confidence and review date.

USER ASSUMPTION

A condition the user chose to hold true for this run. Vigil does not say whether it will occur.

DERIVED CONDITIONAL IMPLICATION

What follows inside the Vigil system if the stated assumptions held — an authored rule citing the record it rests on, never a prediction.

UNKNOWN

Vigil has no reading, or the evidence needed to form one is not recorded.

NO COVERAGE

This series does not measure here. Distinct from a covered period with zero records.

NO MATERIAL CHANGE

The file was reviewed and nothing moved. A finding, not an absence.

SAVED WORK

Your own work. It references canonical records and is private to your account.

Freshness and staleness

Professional inherits the public site's three review bands — Recently reviewed, Review ageing, Review overdue — computed from a record's own review date (120 days current, 270 days ageing). Every Professional record prints the date it is as of, the date it was last reviewed and, where it has one, the date it last changed. There is no composite freshness score and no "updated N hours ago" language: a date is a date. A record whose canonical basis has gone stale carries that band; it is never hidden and never refreshed automatically.

Observable records

What it is. A canonical Watch question names the observables that would bear on it. An observable record extends one of those, by reference, with a dated reading in a stated register (observed, assessed, or unknown), the dated evidence that bears on it and which way, the countervailing evidence Vigil records, its stated limits, its system and Tempo context, and a review ledger of its own.

What it is not. Not a metric, not a threshold, not a probability, not a trend. It cannot mark an observable met — only the canonical Watch record does that, and the two are checked to agree at every build. It cannot restate the question or the condition.

What it uses. Brief items, key events, Analytical Notes, Tempo records and dated references, each named by id and resolved at build. An evidence entry that names no canonical basis fails the build.

How it is maintained. The parent question's weekly review is the check of every observable under it, so "last checked" is the parent's newest ledger date. The record's own ledger is written only when the reading moved, evidence was added, the observable was met or the record was retired. A quiet week costs nothing and is shown as such.

What you may infer. Where the observable stands on the stated date, on the evidence listed. What you must not infer. Whether it will be met, how close it is, or what a "direction" on an evidence entry adds up to — nothing sums directions. When evidence is missing the record says "no evidence recorded" and its reading is labelled unknown.

Professional Standing Watch

The same object as the public demonstration: a selection of canonical Watch questions organised around one exposure, with one authored relevance note per question and a portfolio review state derived from the questions' own ledgers. A Professional scope differs only in being maintained for subscribers and in declaring the sector lenses it serves. It carries no assessment, observable, ledger, source or confidence of its own and is reviewed by the weekly cycle that reviews its questions.

Tempo Lab

Every rule of the Map surface applies unchanged: Tempo is recorded activity, never effect, severity, probability or score. Tempo Lab may place several series on one time axis so their timing can be compared. It never normalises unlike series, never draws a combined line, never derives a percentage change, and always distinguishes a covered period with zero records from a period a series does not cover. A period comparison shows counts per series per window with each series' coverage stated. A chart exported from Tempo Lab carries the inclusion rule, the coverage, the bin size and the sources of every series it shows.

Tempo does not cover every module, by decision. A series exists only where Vigil has a defensible, repeatable inclusion rule it can apply consistently. Current coverage is Ukraine–Russia, Taiwan Strait and Iran–Gulf; the remaining modules maintain no activity series and their Tempo rail is absent rather than empty. A module without Tempo is not a quiet module, is not a less important one, and its absence is never evidence about the level of activity in that theatre. Vigil does not create a series for feature consistency, and the rule for adding one is the rule in the public methodology and nothing weaker.

Evidence Explorer and Query

Evidence Explorer walks the corpus's own references: the ledger entries behind a question, the brief items and events behind those entries, the sources behind those items, and the reverse — where a source is used, what a development bears on. It authors nothing. The Research workspace retrieves canonical records with a stated reason for each match (title, alias, keyword, source, connection) and filters on the records' own metadata. There is no generated answer, no chatbot and no ranking by confidence.

Compare

Comparison dimensions are drawn from the compared records' own fields — an actor's stated and assessed objectives, its constraints by kind, its leverage and red lines; a question's condition, observables and latest outcome; a series' inclusion rule, coverage and counts. Where either object lacks a dimension the cell says not comparable and why. No cell is filled to complete a row, and no similarity is scored.

Scenario Lab

A scenario does not say whether something will happen. It asks: if the stated assumptions held, which parts of the Vigil system become relevant, affected or stressed? The assumptions are the user's and are labelled USER ASSUMPTION. Each implication is an authored rule labelled DERIVED CONDITIONAL IMPLICATION, activated by the chosen assumptions, citing the canonical records it rests on (the VIGIL RECORD), the Watch questions and observables it bears on, and its own unknowns. Its confidence grade is confidence that the dependency exists in the record, never that the outcome occurs.

A template states what it does not model and the conditions under which its premise stops holding. Analogues are named records with stated matches, differences and limits of analogy, never "similar events". Nothing is generated at runtime: the tool shows and greys authored rules. No probability, escalation score, threat level or narrative is produced on any surface, to any reader.

How it is maintained. A template is reviewed monthly and whenever a Watch question it cites changes assessment; the Professional Monitor lists templates whose cited questions moved. A rule whose cited record is stale carries that record's band.

Brief Builder and exports

A briefing is assembled from canonical blocks the subscriber selects. Each block travels with its canonical wording, its dates, its confidence and its sources; the builder can place, order, remove and annotate a block, never edit its text. The subscriber's own paragraphs are labelled as the subscriber's. Exports (Markdown, print, CSV, JSON) carry the same provenance, the canonical URL of every block and the date the export was generated. An automated summary never alters Vigil's analytical meaning because none is produced.

Personal work and privacy

Follows, saved queries, comparisons, scenario runs, brief drafts, timelines and private notes are the subscriber's own and are stored as references to canonical records plus the subscriber's own text. Canonical research is never copied into a personal record; if Vigil corrects a record, every saved reference shows the corrected record. Personalisation changes what a subscriber is shown, never what Vigil assesses. Every personal row is deleted with the account. The privacy notice lists the fields.

Professional Monitor and notifications

The Monitor is derived entirely from the review cycle's own records: material changes, quiet reviews, evidence added, Tempo changed, pending questions and Professional records reviewed. Nothing is written for it separately. A notification, where one exists, is delivery of a reviewed outcome — never detection, prediction or warning — exactly as §10 states.

Sector lenses

A lens — security, energy, maritime, commercial, defence industry, humanitarian, civil resilience, infrastructure, political — is a declared relevance on a scope, an observable record or a scenario template. It changes which records a reader with that exposure meets first. It creates no parallel corpus, no second assessment and no different confidence.

What a subscription does not change

One editorial standard, for everyone. Nothing is assessed differently, weighed differently or stated with more confidence because it is behind a subscription; a correction is issued and logged the same way. Payment buys access to Professional functionality and material, and nothing else.

What Vigil Professional is · Open Professional · Handbook · Terms