write-the-intel-brief

write-the-intel-brief is a report-writing methodology skill for turning investigation findings into a defensible intelligence product that survives challenge. It targets the specific failure that destroys credibility: a sentence that reads as fact but is actually an inference. The core discipline is sorting every claim into observation, inference, or assessment — kept visually distinct — with verb cues (is/lists/shows vs. indicates/implies vs. we assess/we judge) so conclusions are never smuggled into the evidence.

It then requires per-claim sourcing: each retained claim carries its specific source or exact tool query, a UTC retrieval timestamp, a source-reliability grade, and an archived snapshot, with chain-of-custody and SHA-256 hashing for anything that might become evidence. Reports are written BLUF — lead with the key judgements answering the objective question, each as one sentence carrying an estimative term and a confidence level — on the assumption the reader stops after the first half page. A standardized probability ladder (almost certainly, highly likely, likely, roughly even chance, unlikely, highly unlikely, remote) is used consistently and never mixed with raw percentages, and a reference file lists banned vague terms (possibly, may/might/could, suggests) with better replacements. Probability and confidence are treated as two independent axes, both stated with reasons.

It is aimed at analysts and investigators producing regulated compliance reports, due-diligence and investment-committee memos, litigation and disclosure materials, evidence packs, and law-enforcement referrals. This is a purely defensive, methodology-and-template skill about rigor, sourcing, and honest calibration — it generates no offensive capability and touches no sensitive systems, making it benign.