OSINT (Open-Source Intelligence) software refers to tools that collect, aggregate, and analyze publicly available information — from the web, social media, public records, satellite imagery, news, forums, etc. — to produce actionable intelligence, without needing classified or private access.
If you have ever opened an OSINT platform, a threat-intelligence dashboard, or a law-enforcement case management tool and seen a rating like “B2” next to a piece of information, you have run into one of the oldest and most battle-tested scoring systems in intelligence work: the Admiralty Code. Most modern OSINT software — from Maltego to custom-built threat-intel platforms — extends this system with two additional fields: Importance and Confidence. Together, these four fields form the backbone of how analysts decide what to trust, what to prioritize, and what to act on.
This guide breaks down exactly how each field is defined, how they interact, and how you might model them in your own OSINT, fact-checking, or investigative software.
Why These Fields Exist
Raw information is not intelligence. A tweet, a leaked document, a human source, and a satellite image all carry different levels of trust, and even a highly trustworthy source can occasionally pass along bad information. Analysts need a standardized, repeatable way to answer two separate questions before acting on anything:
- Can I trust where this came from? (Reliability)
- Can I trust what it says? (Credibility)
Once those two are scored, two more practical questions follow:
- Does this actually matter right now? (Importance)
- How sure am I, overall, once everything is weighed together? (Confidence)
Separating these concerns prevents a common analytical trap: assuming that because a source is reliable, everything it says must be true — or that because a claim sounds plausible, the source must be trustworthy. Keeping them independent is the whole point of the system.
1. Reliability — Rating the Source
Reliability scores the source itself — a person, sensor, outlet, document, or feed — based on its track record, not on the specific claim being evaluated today. It uses a letter scale from A to F, originally standardized by NATO and still used almost universally in OSINT and military intelligence tooling.
A — Completely reliable : No doubt of authenticity, competency, or trustworthiness.
Long history of total reliability.
B — Usually reliable : Minor doubts. Valid information most of the time.
C — Fairly reliable : Doubts exist. Has provided valid info in the past.
D — Not usually reliable : Significant doubts. Valid info less often than invalid.
E — Unreliable : Lacks authenticity/competency. History of invalid info.
F — Cannot be judged : New or unknown source, no track record to evaluate.Example: A verified government press office with a decade of accurate statements might be rated A. An anonymous, newly created social media account posting about a breaking event, with no prior history, would be rated F — not because it’s necessarily lying, but because there’s simply no basis yet to judge it either way.
2. Credibility — Rating the Information
Credibility scores the specific claim or report, independent of who said it. It asks whether the content is internally logical and whether it lines up with everything else already known. It uses a numeric scale from 1 to 6.
1 — Confirmed : Confirmed by other independent sources; logical; consistent.
2 — Probably true : Not confirmed; logical; consistent with other information.
3 — Possibly true : Not confirmed; reasonably logical; agrees with some info.
4 — Doubtfully true : Not confirmed; possible but illogical; no corroboration.
5 — Improbable : Not confirmed; not logical; contradicted by other info.
6 — Cannot be judged : No basis to evaluate the claim at all.Example: A highly reliable news wire (Reliability A) reporting an unconfirmed casualty figure during a fast-moving event might still only earn a Credibility of 3 (“possibly true”) because the number hasn’t been corroborated yet — giving a combined rating of A3. This is exactly why the two fields are kept separate: a great source can still report an unconfirmed detail.
3. Importance — Rating the Priority
Importance is not part of the original Admiralty Code — it’s a triage layer that most modern OSINT and case-management platforms add on top, because analysts are usually swimming in far more flagged information than they can act on immediately. Importance asks: if this turns out to be true, how much does it actually matter to the investigation, client, or mission?
It’s typically modeled as a simple ordinal scale:
Critical — Directly affects an active decision, safety issue, or deadline.
High — Materially changes the picture of the investigation/target.
Medium — Useful context, adds to the picture but isn't urgent.
Low — Background noise; nice to have, not investigation-critical.Example: A rumor about a company’s internal leadership dispute (Credibility 3) might be rated Low importance for a general due-diligence report, but the exact same claim could be rated Critical if the investigation’s entire purpose is verifying leadership stability before a merger.
4. Confidence — The Analyst’s Overall Judgment
Confidence is different from the other three: it’s not scored per source or per claim, it’s the analyst’s synthesized judgment across an entire assessment, after weighing Reliability, Credibility, corroboration from multiple sources, and analytic reasoning. Most platforms that follow US intelligence community tradecraft standards (ICD 203) express it in three tiers:
High Confidence — Well-corroborated by multiple independent, reliable sources;
strong logical consistency; few alternative explanations.
Moderate Confidence — Credibly sourced but not fully corroborated, or relies
heavily on a single strong source.
Low Confidence — Fragmentary, poorly corroborated, or drawn from
questionable or conflicting sourcing.Key distinction: Credibility rates one report. Confidence rates the analyst’s overall conclusion after looking at everything available — it can go up as more corroborating sources appear, even if no single source individually improves.
How the Four Fields Work Together
Here’s the practical flow an analyst (or an OSINT platform’s scoring engine) typically follows:
Source Reliability (A-F) ---\
>---> Analyst synthesis ---> Confidence (High/Med/Low)
Information Credibility(1-6)--/ ^
|
Corroboration + contradicting evidence
Importance (Critical/High/Medium/Low) = a SEPARATE, independent axis
set by the investigator based on relevance to the case, not derived
from Reliability/Credibility/Confidence.Many platforms then plot Importance against Confidence as a 2×2 or 4×4 triage matrix, which is one of the most useful visual tools in OSINT work:
| High Confidence | Low Confidence | |
|---|---|---|
| Critical Importance | Act on it now | Urgent verification needed — do not act yet |
| Low Importance | Log and move on | Deprioritize / archive |
This is exactly why “High Importance + Low Confidence” is one of the most dangerous combinations in any investigation — it’s tempting to act because it matters so much, but the evidence isn’t there yet. Good OSINT software flags this combination visually rather than letting it blend in with everything else.
Worked Example
Imagine an OSINT platform ingesting three reports about a single event — a factory fire:
- Report 1 — Verified fire department scanner feed, direct observation. Reliability: A. Credibility: 1 (confirmed). Importance: High. Confidence: High.
- Report 2 — Unverified eyewitness post on social media, no prior history. Reliability: F. Credibility: 3 (possibly true, matches Report 1). Importance: Medium. Confidence: Moderate (because it’s corroborated by an A1 source even though the source itself is unrated).
- Report 3 — Anonymous forum post claiming the fire was arson, contradicted by the fire department’s statement. Reliability: E. Credibility: 5 (improbable, contradicted). Importance: Critical (if true, it changes the nature of the case) but Confidence: Low — meaning: flag it, don’t act on it yet.
Key Takeaways
- Reliability rates the source’s track record — not today’s claim.
- Credibility rates today’s claim — not the source’s history.
- Importance is an independent triage axis based on relevance and urgency, not truthfulness.
- Confidence is the analyst’s synthesized judgment across the whole body of evidence, and can shift as corroboration changes even when individual scores don’t.
- The most useful thing OSINT software can do is keep these four fields visually and structurally separate, then surface the dangerous combinations — especially high importance paired with low confidence — rather than collapsing everything into a single “trust score.”