UK Marking - OFFICIAL / OFFICIAL-SENSITIVE
Detects the UK Government Security Classifications Policy (GSCP) OFFICIAL classification and its -SENSITIVE marking, written OFFICIAL-SENSITIVE and applied immediately after the OFFICIAL classification per the current policy. Matches the verified structured "CLASSIFICATION - HANDLING INSTRUCTION[- DESCRIPTOR]" format (e.g. OFFICIAL-SENSITIVE - RECIPIENTS ONLY, OFFICIAL-SENSITIVE - COMMERCIAL) using the centrally-defined GSCP handling-instruction and descriptor vocabulary, the bare OFFICIAL-SENSITIVE banner, and — at a lower, UK-government-context gated tier — bare OFFICIAL alone, which is a common English word. Regex logic verified directly against the current Government Security Classifications Policy (gov.uk, version 2.0, August 2024); matched case-sensitively so ordinary lowercase English use of "official"/"official-sensitive" does not trigger it.
- Type
- regex
- Engine
- boost_regex
- Confidence
- high
- Confidence justification
- High confidence on the OFFICIAL-SENSITIVE hyphenated marking, in both its structured handling-instruction/descriptor form and its bare banner form: both are case-sensitive, structurally distinctive tokens verified directly against the current GSCP primary source, with negligible natural-English collision risk. Bare OFFICIAL alone (65) is the widest tier and is highly collision-prone — a common English adjective used constantly outside government contexts (OFFICIAL WEBSITE, OFFICIAL SPONSOR, OFFICIAL RECEIVER, OFFICIAL SECRETS ACT citations, etc) — so it requires UK central-government/GSCP corroborative evidence within 300 characters plus a dedicated NOT-exclusion for the most common non-marking collocations.
- Jurisdictions
- uk
- Regulations
- Official Secrets Act 1989
- Frameworks
- GSCP
- Data categories
- government, security-classification
- Scope
- narrow
- Risk rating
- 6
- Platform compatibility
- Purview: Compatible, GCP DLP: Compatible, Macie: Compatible, Zscaler: Compatible, Palo Alto: Compatible, Netskope: Compatible
Pattern
\bOFFICIAL[\s-]+SENSITIVE\bCorroborative evidence keywords
HM Government, Cabinet Office, GSCP, protective marking
Proximity: 300 characters
Should match
This document is marked OFFICIAL-SENSITIVE - RECIPIENTS ONLY and must not be forwarded outside the distribution list.— Structured handling-instruction form per the verified GSCP "CLASSIFICATION - HANDLING INSTRUCTION" formatBriefing note: OFFICIAL-SENSITIVE - COMMERCIAL — do not release without clearance from the commercial team.— Structured descriptor form per the verified GSCP "CLASSIFICATION - HANDLING INSTRUCTION - DESCRIPTOR" formatThe attached report is classified OFFICIAL-SENSITIVE.— Bare OFFICIAL-SENSITIVE banner, no descriptor requiredCover sheet reads OFFICIAL- SENSITIVE for internal circulation only.— PDF line-wrap between the OFFICIAL- hyphen and SENSITIVE (the [\s-]+ separator tolerates a wrap-inserted newline)This memo must be handled under Cabinet Office guidance; it is marked OFFICIAL and should be treated accordingly by HMG staff.— Gated bare-OFFICIAL form — UK central-government/GSCP corroborative evidence (Cabinet Office, HMG) present within 300 characters
Should not match
please visit our official website for further official announcements— lowercase English word (case-sensitive exclusion)This memo is UNOFFICIAL and carries no protective marking.— UNOFFICIAL must not match OFFICIAL (no word boundary before OFFICIAL within UNOFFICIAL)OFFICIAL SPONSOR OF THE NATIONAL SPORTS CHAMPIONSHIP — MARKETING PARTNERSHIP ANNOUNCEMENT— ALL-CAPS non-government collocation — excluded from the top-level pattern (no -SENSITIVE present) and from the gated bare tier by the OFFICIAL-collocation NOT-groupDISCLOSURE OF THIS MATERIAL MAY CONSTITUTE AN OFFENCE UNDER THE OFFICIAL SECRETS ACT 1989.— ALL-CAPS statutory citation — excluded from the top-level pattern (no -SENSITIVE present) and from the gated bare tier by the OFFICIAL SECRETS ACT NOT-group
Known false positives
- ALL-CAPS prose (legal boilerplate, headers, or documents rendered entirely in capitals) defeats the case-sensitivity signal the bare OFFICIAL-SENSITIVE banner tier relies on — any all-caps rendering of the phrase can fire with no protective-marking intent. Mitigation: Treat banner-tier-only (85) hits as lower confidence when the surrounding text is itself all-caps; the structured descriptor tier (85) with a verified GSCP handling-instruction/descriptor term remains the more reliable signal.
- OFFICIAL is an extremely common English adjective in everyday non-government usage (OFFICIAL WEBSITE, OFFICIAL STATEMENT, OFFICIAL SPONSOR, OFFICIAL PARTNER, OFFICIAL RECEIVER, OFFICIAL LAUNCH/OPENING, and citations of the OFFICIAL SECRETS ACT itself), all more common in general text than the GSCP classification marking. Mitigation: The bare tier (65) requires UK central-government/GSCP corroborative evidence and excludes this collocation set via a dedicated NOT-group. The structured and banner OFFICIAL-SENSITIVE tiers (both 85) are unaffected — they are structurally distinctive and do not require the gate.
- This SIT's OFFICIAL/OFFICIAL-SENSITIVE vocabulary overlaps with the Australian PSPF scheme: AU's au-marking-official ships its own bare OFFICIAL banner tier (85) ungated apart from a noise-exclusion NOT-group (no AU-context requirement), and au-marking-sensitive's OFFICIAL: Sensitive stem tier (90) uses a different separator convention (colon/space/`//`, not a fixed hyphen). A genuine UK OFFICIAL or OFFICIAL-SENSITIVE document will typically NOT co-fire the AU SITs (different separator syntax), but the bare word "OFFICIAL" appearing in UK government prose could coincidentally satisfy AU's ungated bare-OFFICIAL tier if it also happens to read as an AU marking banner. Mitigation: Deploy policies should expect occasional cross-jurisdiction co-fire on the bare-word vocabulary as a corroborating, not conflicting, signal; do not add a UK-exclusion to the AU SIT, as that would break its own (broader, deliberately ungated) intended coverage.
- The GSCP's "[INSERT ORGANISATION(S) NAME] USE ONLY" handling instruction and centrally-allocated Codewords are not machine-matchable: organisation names and codewords are arbitrary text assigned per-asset, not fixed vocabulary, so no regex can enumerate them. Similarly, the GSCP's REL EU releasability marking (used alongside a UK-prefixed OFFICIAL-SENSITIVE asset shared with an EU institution) is rendered inconsistently in the primary source itself (prose text reads "REL EU", the source's own worked example renders it "REL-EU" on the following line) — the exact character-level syntax cannot be pinned down from the primary source, so it is not implemented. Mitigation: No mitigation implemented — documented as a known coverage gap rather than guessing at unverified syntax. The structured handling-instruction/descriptor tier and the bare banner tier remain the reliable, verified signals.