OSINT Analyst Course

module-card-template

8 min read

Module Card Template

Every module card follows this shape. The card is the authoring artifact; the XMind Pitch deck is a downstream view, so keep the teaching body outline-shaped (clean heading + bullet hierarchy) for clean import. Delete sections that genuinely don’t apply; don’t pad — depth scales the card (see Depth scaling).

The card teaches by structured inquiry, not assertion — Heuer/Pherson SAT discipline, not corporate L&D. The Socratic scaffolding below (hinge, calibration prompt, > Q: checks, pre-mortem, self-critique exercise, diagnostic-question errors, adversarial questions) is doctrinal, not decorative. It is what forces System 2 to engage; cut it and the card recites tradecraft while teaching anti-recitation.

Copy everything in the fenced block into <ID>-<slug>.md (the filename stem from course-architecture.md §8).

---
id: <F1-01>
title: <Two minds — System 1 / System 2, WYSIATI>
shelf: <F1>
depth: <n/a | thin | med | deep>      # n/a for foundations/core; tier for applied cards
status: <planned | drafting | review | done>
prerequisites: [<L1-01>]              # card IDs a student needs first; [] for the spine's entry card
serves: [<OSINT-029>]                 # APPLIED cards only — deliverable IDs. Omit on core/foundations
                                      #   (a core card's consumers are DERIVED from applied prerequisites)
sources: [<kahneman>]                 # source IDs from the six-book canon (see _provenance/sources.md)
tags: [<cognitive-bias>, <calibration>]  # cross-cutting themes — MUST be drawn from the tag registry (§9)
last-verified: <YYYY-MM-DD>           # REQUIRED for any LIVE/tooling card; omit for stable doctrine
---
 
# <ID> · <Module Title>
 
## Learning objectives
- **Terminal:** after this module the analyst can <observable capability>.
- **Enabling:** <2–4 sub-skills that build to the terminal objective>.
 
## Concept
 
**From <prereq-ID>:** you carry <what's inherited>; this card adds <what's new>.
<!-- The hinge. One line, names the bridge — NOT the answer. Omit on the spine's entry card. -->
 
> **Calibrate:** <a question the analyst attempts cold, before any exposition>. Hold your answer.
<!-- Forces System 2 before reading; we return to it. The hinge must not pre-answer this. -->
 
<Exposition — outline-shaped (headings + bullets) for clean XMind import.
Sources are the frontmatter `sources:` list. Where a student would want a deep dive, add a light
chapter pointer in prose (e.g. "Kahneman Ch. 21") — not a claim-by-claim citation apparatus.
Full ICD-206 product rigor is **taught as content** in L5-03 and required of student *products*, not of cards.
At genuine inference forks — where intuition misleads — drop a Socratic check (predict, THEN read on).
2–4 per card, on real forks, not every sentence: >
 
> **Q:** <prediction-forcing question at an inference fork>.
 
*Nudge:* <when the Q asks the student to do or produce something, follow it with one or two italic
sentences: a light example or partial instance from the card's own case that gives a foothold without
answering the Q (operator rule, 2026-07-30)>.
 
## Procedure (if the module teaches a method)
<Numbered steps. If doctrine disagrees on the steps (e.g. ACH Heuer-8 vs SAT-9), give the canonical sequence
here and note variants under "Doctrinal notes.">
 
**Pre-mortem.** Imagine you ran this and the analysis still failed. What most likely went wrong, and
at which step? <Answered before moving on — a red-team without naming it.>
 
## Worked example
<One concrete pass through the concept/procedure on realistic OSINT data — ideally a scenario that maps
to a deliverable this module serves. Code blocks ONLY where the procedure executes as code (query syntax,
regex, a transform); otherwise narrate the reasoning on the data.>
 
## Practical exercise — the competency gate, not enrichment
<A do-it task with a concrete deliverable. The analytic discipline is built in, not wrapped around:>
1. Produce the judgment / product.
2. Write the strongest critique of your own judgment.
3. Revise in light of it.
<Self-check / rubric: competency is demonstrated, not quizzed.>
 
## Common errors
<The 3–5 ways analysts get this wrong. Each as Tell → Diagnostic question → Fix — the diagnostic
question is what an analyst runs on live casework at 0200; naming the bias is not.>
 
| Tell (what you'd observe) | Diagnostic question | Fix |
|---|---|---|
| … | … | … |
 
## Figures
<Match the visual to the content (see *Visual decision rule*). Operator verifies every diagram.
Doctrine-figure recreations may note the source book/chapter (e.g. "after Heuer Fig 15 — ACH matrix").>
 
## Doctrinal notes / variants
<Where sources disagree, or house standard overrides — name the variant and the call. When the card
teaches a CHOICE among ≥2 methods, use a comparison table, not narrative pros/cons:>
 
| Method | When to use | Cost | Failure mode | Confidence ceiling |
|---|---|---|---|---|
| … | … | … | … | … |
 
## Adversarial questions
<4–6 red-team prompts a senior analyst would push on this card's content / the analyst's would-be
product. Doubles as the operator's review question bank.>
-
 
## Links
- Prerequisite cards: <IDs>
- Composed by applied tracks: <A–G IDs>     (core cards only)
- Deliverable(s): /home/farseer/Projects/osint-cell-dev/deliverables/<type>/OSINT-NNN-*.md
- Collection map(s): /home/farseer/Projects/osint-cell-dev/collections/<type>/*-collection.md
 
## Sources
<source-id(s) from frontmatter; optional one-line chapter pointers, e.g. "Kahneman Ch. 1–3 (two systems); Ch. 7 (WYSIATI)">
 
**Next:** [[<next-ID-stem>|<next-ID>]]
<!-- Forward-pointer. Omit on a track's terminal card (or point to the capstone). -->

Notes

IDs & links

  • IDs are stable. Once a card has an ID, cross-links depend on it — don’t renumber casually.
  • Wikilink by full stem, ID as piped label ([[F1-01-two-minds|F1-01]]); register the stem in course-architecture.md §8 before linking. Frontmatter only for the metadata above — no duplicate visible bullets; Obsidian’s Properties panel renders it, and DRY means status can’t disagree with itself.
  • Cross-link, never copy. If you’re re-explaining a core primitive in an applied card, stop and link it.

Frontmatter schema (linter keeps keys sorted)

  • serves lives on applied cards only (deliverable IDs — genuinely authoritative). A core card’s consumers are derived from applied prerequisites via Dataview, not stored by hand.
  • tags MUST be drawn from the tag registry (course-architecture.md §9) — free-text tags rot the inverse index. sources is the queryable list of canon source IDs for the card.

Voice register — IC analytic

  • Declarative, source-anchored, no hype. Models: Heuer Psychology of Intelligence Analysis, Kent, the six-book civilian canon. Anti-models: Substack tech blogs, corporate L&D, pop-pedagogy.
  • No hype adjectives — cut powerful, fascinating, incredible, superhuman, magic, transformative, revolutionary, game-changing, seamless, world-class. No decorative epigraphs; quote only where the line carries the argument.
  • Humanizer pass (house rule, 2026-07-29). All card prose gets a humanizer pass before verify — strip AI-writing tells (inflated symbolism, vague attribution, rule-of-three padding, filler transitions, negative parallelisms). No em dashes in card prose — hard gate, not a flag; restructure the sentence or use a colon, comma, or period instead.

The Socratic hooks operate at different altitudes — they are not redundant:

  • > Q: — inline comprehension/inference, mid-Concept.
  • Pre-mortem — failure of the procedure, end of Procedure.
  • Self-critique — the analyst’s own judgment, inside the exercise.
  • Tell → Diagnostic question → Fix — portability to live casework, in Common errors.
  • Adversarial questions — red-team of the product as a whole, its own section.

Visual decision rule

  • Markdown table — tabular / comparison data (renders natively everywhere). Default for anything tabular.
  • Mermaid — any actual diagram: branching, layers, state, networks, decision trees, multi-actor sequences. Doctrine-figure recreations may note the source book/chapter.
  • Prose — interpretive content a diagram would distort (e.g. a tradeoff “curve”).
  • ASCII only as a rare, operator’s-call exception when something is genuinely clearer as monospace.

Depth scaling

  • Foundations / core (F, L): full card — the full Socratic apparatus is required here; this is where thinking is taught.
  • Thin applied: composes core, doesn’t re-teach. Keep the calibration prompt, ≥1 > Q:, the self-critique exercise, and adversarial questions; the Procedure pre-mortem applies only if the card teaches a procedure.

Sources (light convention) Cards list canon IDs in frontmatter sources: and may add optional chapter pointers for deep dives (“Kahneman Ch. 21”). They do not carry a claim-by-claim ICD-206 provenance table. Full ICD-206 rigor is course content in L5-03 and is required of student products, not of teaching cards.

Verifier gates (what the coursebuild verifier checks — keep this template and that prompt in sync): schema conformance (keys ⊆ schema · tags ⊆ registry · applied cards carry serves deliverables) · ≥1 calibration prompt · ≥1 > Q: · every do/produce > Q: followed by an italic *Nudge:* line (light example, does not answer the Q) · Procedure pre-mortem if a Procedure exists · the 3-step self-critique exercise · Common errors as Tell → Diagnostic → Fix · Adversarial questions present (4–6) · comparison table when ≥2 variants are taught as a choice · Mermaid for diagrams (not ASCII) · hype-adjective grep (flag for review) · no decorative epigraphs · cross-link-not-copy · wikilink form (full §8 stem, ID as pipe label) · no em dashes in card prose (hard gate) · humanizer pass done (AI-tell scrub) before verify.

Graph View