Editorial Standards

Intelligence writing lives or dies on a single discipline: the clear separation of confirmed fact from analytical assessment. These are not interchangeable. Conflating them — presenting an assessment as established fact, or burying confirmed evidence inside hedged language — destroys the analytical credibility that is SpyWitness's primary editorial asset.

Every piece published on SpyWitness is written to maintain that distinction at every sentence. Readers should always know whether they are reading confirmed reporting, assessed probability, or analytical opinion. The publication uses calibrated language to make this explicit — not as a stylistic preference, but as a non-negotiable editorial standard.

SpyWitness applies a five-level confidence framework to all claims, consistent with standard intelligence assessment practice.

The language is mandatory. No claim is presented without a calibration level — either explicit in the text or determinable from context. Where confidence cannot be established, the claim is held or flagged as unverified.

Named sources are always preferred. Official sources — government statements, court documents, parliamentary records, declassified material — carry the highest evidentiary weight and are cited directly. OSINT-derived claims name the data type and, where possible, the platform or methodology: satellite imagery, flight-tracking data, financial filings, intercepted communications reported by credible outlets.

Anonymous sources are used where operational security justifies it and the source's access to the reported information is assessed as credible. Anonymous sourcing is always attributed with the minimal available characterisation — "a serving official," "a source with direct knowledge of the operation" — and never presented as authoritative without corroboration from at least one additional independent source. Single-source claims are flagged as such.

SpyWitness does not publish unsourced claims. Where a claim is made by a named actor — a government, intelligence service, militia, or company — it is attributed to that actor, not adopted as independently verified fact unless corroborated. The distinction between what an actor claims and what the evidence supports is maintained in every piece.

Secondary sourcing is attributed to the originating outlet. SpyWitness does not re-present second-hand reporting as its own primary source material.

Open-source intelligence reporting at SpyWitness makes its methodology visible. Readers are told what data was collected, from what sources, and by what analytical method it was assessed. Transparency is the credibility of OSINT work — conclusions derived from satellite imagery, vessel-tracking systems, or commercial data aggregators are only as strong as the methodology behind them, and that methodology is disclosed.

Errors are corrected promptly and transparently. Where a material factual error has been published, a correction is appended to the relevant piece noting what was incorrect and what the accurate information is. Corrections are not deleted from the record — the original error and its correction are documented. Editorial judgments and assessments that prove incorrect over time are not classified as errors requiring correction; they are acknowledged in subsequent analytical updates.

Correction requests can be submitted via the contact page. All requests are reviewed by the editorial desk.

SpyWitness does not publish unverified allegations as news. It does not sensationalise threat assessments to generate traffic. It does not present analytical opinion as confirmed intelligence. It does not suppress sourcing limitations to present a cleaner narrative. It does not act as a conduit for government or institutional messaging without transparent attribution.

These are not aspirational commitments. They are operational constraints applied in the editorial process before publication.