Editorial Methodology

The Signal
Confidence
Index.

Every signal IN-KluSo publishes is scored before it goes live. The SCI is the methodology behind that score — five dimensions, one publication threshold, no shortcuts.

5
scoring dimensions, each measuring a different kind of rigor
0.65
minimum composite score to publish — below this, it doesn't run
0.976
highest SCI score on record — CORE-002: The 5x Trap
The Problem We're Solving

Most intelligence fails
before it reaches you.

Not because the analysis is wrong — because the analysis was never grounded in the first place. Sources weren't named. Mechanisms were implied. Alternatives were ignored. Uncertainty was buried. The SCI exists to make those failures visible and preventable, before anything is published.

The Five Dimensions

What we score,
and why it matters.

Each dimension targets a specific failure mode in intelligence reporting. A signal can score perfectly on four and collapse on one — which is why all five are required.

EVID 35%
Evidence Quality
The weight of the claim is only as strong as what's behind it. EVID scores the quality and independence of sources — primary documents, institutional data, and verified records carry the most weight. Commentary and opinion carry almost none. This dimension carries the highest weight in the composite score because without it, everything else is speculation presented as analysis.
Who said this? Can it be verified independently? Is this primary or derived? What's the source's institutional standing?
MECH 25%
Mechanism Clarity
The difference between a signal and a headline is a mechanism. A headline tells you what happened. A signal tells you why — with a named, traceable causal chain from event to consequence. MECH scores whether that chain is complete, evidenced, and free of logical gaps. Partial chains score partially. Descriptions of outcomes with no mechanism score near zero.
Why did this happen? What caused what? Is the causal chain complete? Can the mechanism be named?
LENS 30%
Lens Coverage
Every signal is examined through three lenses: epistemological (how do we know what we know), systems (what structural forces are at play), and behavioral (how are actors actually responding). The weight of each lens shifts by division — CORE weights epistemology and systems equally; GROUND weights behavioral observation more heavily. A signal that sees through only one lens is incomplete by design.
How was this knowledge produced? What system is operating here? What are people actually doing?
RISK —
Uncertainty Disclosure
Every analysis has limits. RISK scores how honestly those limits are named. It is not a penalty — a signal that clearly separates what is verified from what is inferred, and explicitly states what it cannot confirm, scores better than one that implies false certainty. IN-KluSo treats uncertainty as editorial integrity, not weakness.
What does this analysis not know? What is verified vs. inferred? What would change the conclusion? Are alternative explanations presented?
TERR 10%
Territory Specificity
Signals without territory are press releases. TERR scores the geographic and temporal precision of a piece — city named, time window stated, actors identified, observable behavior documented. Vague geographic language ("urban areas," "major metros") scores near zero. A street, a permit, a month, a named institution — those score at the top.
Which city? Which neighborhood? When — month and year at minimum? Who is the named actor or institution? What behavior was observed directly?
Evidence Quality — Source Tiers

Not all sources
are created equal.

The EVID dimension runs on a three-tier source hierarchy. Tier determines how much a source contributes to the composite score. No exceptions.

Tier A
Primary Sources Institutional reports, regulatory filings, court records, peer-reviewed research, government data, company disclosures. Load-bearing claims require at least one Tier A source.
Full weight
Tier B
Secondary Sources Major publications accurately citing primary data with clear methodology. Acceptable for numeric claims when Tier A is unavailable — never as the only citation.
60% weight
Tier C
Commentary Blogs, newsletters, LinkedIn posts, opinion pieces, anonymous sources. Used for color and context only. Never cited for numeric or load-bearing claims.
Color only
HARD CAP

Zero Tier A sources and fewer than two independent Tier B sources → SCI ceiling of 0.55. No signal can achieve HIGH confidence without primary sourcing. This rule has no override.

Publication Tiers

What the score
determines.

The SCI composite produces a confidence tier. The tier determines whether and how a signal publishes. There is no editorial override.

0.80–1.00
HIGH
Publishes as anchor signal. Flagged in division index. Available for audio and distribution.
0.60–0.79
MODERATE
Publishes with explicit uncertainty disclosure. Gaps noted in-line. Monitoring recommended.
0.40–0.59
LOW
CRUMB only. Published as a minimum unit. Requires additional sourcing before full signal treatment.
0.00–0.39
MINIMAL
Does not publish. Archived for monitoring. Returns if new primary sourcing becomes available.
In Practice

60 signals.
51 HIGH confidence.
The system works.

Every published signal in the IN-KluSo archive carries its SCI score at the top of the article. The methodology is not a marketing claim — it's the filter everything passes through before it reaches you.

60
Published signals
85%
HIGH confidence rate
~0.90
Average SCI score
0.976
Record — The 5x Trap
↗ HumanKlu™

The same standard,
applied to your content.

HumanKlu™ is the external certification standard built on the same five SCI dimensions. Organizations submit content for scoring — and certified content earns the HumanKlu™ badge: a visible, verifiable signal that the work was produced to a documented human standard, not generated and published unchecked.

EVID MECH LENS RISK TERR

Discussion