StandardGlobalTradecraftHigh Confidence

Assessment Style Guide

datetime and versioning standard, structured metadata standard, explainer standard, and accumulated practice across the assessment corpus.

Review Queue Full Index

UNCLASSIFIED//OPEN SOURCE

Standard ID: WI-STD-003

Prepared UTC: 2026-06-19T00:00:00Z

Information cutoff UTC: 2026-06-19T00:00:00Z

Source base: WARLOCK-INDEX product standard, source evaluation standard, datetime and versioning standard, structured metadata standard, explainer standard, and accumulated practice across the assessment corpus.

Analytic confidence: High for internal style guidance and repository writing requirements.

Topics: tradecraft; style; writing standards; assessments Source classes: Internal standard; analytic tradecraft Safety boundary: Style and voice guidance only. Does not authorize policy recommendations, operational detail, targeting, or any prohibited content.

Voice

Write in clear, disciplined prose. The tone should be sober, direct, and human: more like a serious defense assessment than a generic report.

Use active sentences where possible. Avoid slogans, hype, filler, and theatrical language. Explain why a fact matters, but do not drift into policy advice.

Structure

Lead with the bottom line. Then present key judgments, evidence, analysis, indicators, and source base.

Separate:

  • What is known.
  • What is assessed.
  • What is uncertain.
  • What would change the assessment.

No Recommendations

Do not write "the United States should" or equivalent prescriptive language in assessment products. Preferred alternatives:

  • "Decision relevance:"
  • "Strategic significance:"
  • "Implication for U.S. interests:"
  • "Indicator to monitor:"
  • "Information gap:"

Treatment Of Adversaries

Use precise actor names. Avoid caricature. Identify stated objectives, capabilities, constraints, and behavior. Strategic clarity is more useful than emotional language.

Dates

Use UTC timestamps for product metadata. Use absolute dates in the body when describing events. Avoid unsupported phrases such as "recently" unless the date is also provided.

Filename format: YYYY-MM-DDTHHMMZ-short-descriptive-title.md

Header Fields (Assessments)

Assessments use this header style:

**Classification:** UNCLASSIFIED//OPEN SOURCE
**Handling:** Public open-source research
**Product ID:** WI-ASMT-AREA-YEAR-NNNN
**Prepared UTC:** YYYY-MM-DDTHH:MM:SSZ
**Information cutoff UTC:** YYYY-MM-DDTHH:MM:SSZ
**Scope:** ...
**Exclusions:** ...
**Source base:** ...
**Analytic confidence:** ...

Other product types (explainers, source packets, trackers, source notes, actor profiles, maps) use a bare classification line followed by a typed ID:

**UNCLASSIFIED//OPEN SOURCE**
**Explainer ID:** WI-EXPLAINER-...
**Prepared UTC:** ...

See product-standard.md and the relevant product-type standard for exact requirements. Always include Prepared UTC and Information cutoff UTC.

Boundaries And Safety Language

Every product must state its boundary explicitly. Use the language in the relevant template and product standard.

Common required exclusions:

  • No recommendations for government, military, or intelligence action.
  • No target selection, weapons employment, or tactical procedures.
  • No exploit steps, evasion methods, or cyber attack instructions.
  • No facility vulnerability mapping or operational planning support.

State the boundary early (often in the header block and again in a dedicated section or the Bottom Line area).

Source Attribution In Text

  • Cite by publisher, document title or series, and date.
  • Distinguish official statements from analysis.
  • Use short quotes when wording is analytically important; otherwise summarize.
  • Record full source details in the Source Base section and in the linked source packet when one exists.

Estimative Language

Use consistent probability language (see also product-standard.md):

  • Almost certainly: very high likelihood.
  • Likely: strong likelihood.
  • Probably: more likely than not.
  • May or could: plausible but uncertain.
  • Unclear: evidence is insufficient or contradictory.

Avoid false precision. Do not assign numeric probabilities unless the source itself provides a defensible basis.

Revision Practice

Do not silently overwrite major judgments. For material changes create a new dated product and link the prior version. For minor corrections, update in place and add a short revision note at the end or in a dedicated section.

See datetime-and-versioning.md and product-standard.md for more detail.

Use the Standards hub, Collections hub, and Documentation Index for navigation and related material.

Cross References