Tools · Attack surface inventory · 16 / 23
subfinder: a source count is not a confidence score
The sources beside a hostname are a list of datasets that returned it. The length of that list is not a liveness score and not a severity. Write the names, not a rank you invented.
The brackets are a list
A line can show a hostname and, in brackets, the sources that reported it on this run. Three names in the brackets means three datasets returned the hostname. It does not mean three live checks, and it does not mean the name is three times as important as a name with one source.
A single source is still a report from that source. It is not a weak rumor you may drop without saying you dropped it, and it is not proof the name is current. The line does not include the date the source first saw the name.
Do not turn the count into a score
Sorting the inventory by how many sources appear is a choice you document. It is not a field subfinder computed as confidence. If you keep only names with two or more sources, write that filter. The names you removed are still in the raw output.
Do not write “confirmed” because the list is long. Confirmation would be a resolution or an owner, cited separately.
What to file
File the hostname, the source names, and the run time. If you mention a count, show the list it came from. Do not file a 1-to-5 score that is not on the line.
Worked case: three sources called confirmed
labs-status.hackblue-lab.org is printed with crt.sh, certspotter, and anubis. A draft marks it “confidence high” and adds it to the live web inventory. No resolution was run. The line has no dates.
The note lists the three source names and the run time. It says the count is the length of that list, not a confidence score, and the name was not resolved in this step. The live inventory does not gain the host from the count alone.
host: labs-status.hackblue-lab.org sources: crt.sh, certspotter, anubis count: 3, which is the length of that list resolution: not run not: confidence high, and not a live web host
Diagram
A source list and a score
The list is on the line. The score is not.
Sources. Dataset names that returned the host
Count. How many names are in the brackets, if you count them
Missing. A date the source saw the name
Confidence. A rank you would have to invent
Liveness. No resolution
Severity. No incident scale
Check yourself
1. Three sources are printed beside one hostname. What is three?
2. A name has one source. What may you do without writing a filter?
3. Why not write “confirmed” from a long source list?