Make approach to defining enums consistent across the project

#654 introduces the enum type Ecosystem implemented differently to other enums (e.g. analysis.Mode and staticanalysis.Task and analysisrun.DynamicPhase)

We should develop a consistent approach to how we define enums:

Some areas that should be covered:

• should the base type be string vs numeric
• how do we handle lists of enums
• how do enums interact with flags
• how do enums handle serialization (JSON)

Both:

• require String()

Numeric based enums:

• can use iota
• can consume as little as 8bits (byte)
• very fast to compare equality
• encoding.TextMarshaler and encoding.TextUnmarshaler needed to support flag.TextVar and json serialization
• requires strings to be specified for marshaling and unmarshaling (although an array of strings may make this easier)

String based enums:

• consume 128 bits per enum
• simpler String() function and serialization
• equality is more costly (at minimum two comparisons, maximum the size of the shortest string)

Compare the enum definitions introduced by #654 for Ecosystem with analysis.Mode, staticanalysis.Task, and analysisrun.DynamicPhase. Review how each handles strings, lists, flags, JSON serialization, and String(), then agree on a single documented approach covering numeric versus string bases and required marshaling behavior.