Skip to content

Spec coverage

Coverage is a weighted measure of how confirmed the corpus is, computed over every attribute and child across all stanzas:

coverage% = (confirmed + 0.5 × probable) / total

It is a progress signal for the research speedrun, not a quality score. A low number is expected and honest while captures are still scarce.

Overall coverage: 25% across 112 documented facts.

Overall

Scope Confirmed Probable Speculative Unknown Total Coverage
All facts 0 57 55 0 112 25%

By category

Category Confirmed Probable Speculative Unknown Total Coverage
media 0 6 5 0 11 27%
signaling 0 43 31 0 74 29%
transport 0 8 19 0 27 15%

By technique

Each row counts the attributes whose provenance cites that technique (an attribute confirmed by several techniques is counted under each).

Technique Confirmed Probable Speculative Unknown Total Coverage
websocket-capture 0 46 36 0 82 28%
baileys-instrumentation 0 11 2 0 13 42%
frida-hooking 0 1 3 0 4 13%
mitm-tls 0 0 0 0 0 0%
static-smali-analysis 0 1 1 0 2 25%
memory-dump 0 0 0 0 0 0%
wasm-analysis 0 0 0 0 0 0%
(no technique) 0 0 0 0 0 0%

Flavor corroboration

Independent reimplementations (flavors) that reproduce a fact. This is not part of the coverage score above — a flavor is a corroborating source, not a technique. Counts use derives_from, so a port and its upstream count once.

  • Facts citing at least one flavor: 3
  • Facts with two or more independent flavors: 0

Back to spec overview