agentLabelFilter is useful when you want to scope a rule to a single client process or browser. It is one of the few filters that can look beyond the HTTP layer thanks to Fluxzy's process resolution, which makes it well suited to multi tenant captures.
Use agentLabelFilter when you want to act on traffic depending on which agent emitted it. The agent label is the string Fluxzy resolved for the source of the exchange. On a transparent capture this is often the process name, in a regular proxy connection it usually mirrors the User-Agent header.
Typical situations:
The filter fires on the requestHeaderReceivedFromClient scope. When you specifically want to match on the literal User-Agent header value, requestHeaderFilter with headerName: User-Agent is equally valid and more explicit.
An updater process from a third party tool keeps polling and saturating your capture. Throttle just that agent so you can keep working on the actual issue.
rules:
- filter:
typeKind: AgentLabelFilter
pattern: updater.exe
operation: Contains
actions:
- typeKind: AverageThrottleAction
bandwidthBytesPerSeconds: 8000
Tag every exchange emitted by Firefox so you can isolate it from Chromium traffic when reviewing the capture afterwards.
rules:
- filter:
typeKind: AgentLabelFilter
pattern: Firefox
operation: Contains
actions:
- typeKind: ApplyTagAction
tag:
value: source-firefox
Reject any request that comes from a known headless runner used for unrelated test suites, so it does not pollute your active investigation.
rules:
- filter:
typeKind: AgentLabelFilter
pattern: ^(HeadlessChrome|playwright|puppeteer)$
operation: Regex
caseSensitive: false
actions:
- typeKind: RejectAction
Select exchanges according to configured source agent (user agent or process) with a regular string search.
Evaluation scope defines the timing where this filter will be applied.
requestHeaderReceivedFromClient This scope occurs the moment fluxzy parsed the request header receiveid from client
agentLabelFilter
The following table describes the customizable properties available for this filter:
| Property | Type | Description | DefaultValue |
|---|---|---|---|
| pattern | string | The string pattern to search | |
| operation | exact | contains | startsWith | endsWith | regex | The search operation performed | contains |
| caseSensitive | boolean | true if the Search should be case sensitive | false |
| inverted | boolean | Negate the filter result | false |
The following examples apply a comment to the filtered exchange
Retains only exchanges with the exact agent label.
rules:
- filter:
typeKind: AgentLabelFilter
pattern: Chrome
operation: Contains
actions:
- typeKind: ApplyCommentAction
comment: filter was applied
View definition of AgentLabelFilter for .NET integration.
This filter has no related filter
Fluxzy fills the label from the most reliable source available. On Windows and macOS it can be the originating process when transparent capture is on, otherwise it falls back to the User-Agent header.
No by default. Set caseSensitive to true if you need to distinguish between agents that differ only in casing.
requestHeaderFilter looks at one raw HTTP header, while agentLabelFilter looks at the resolved agent label which may come from outside the HTTP request, for example from the process table during transparent capture.
Yes. Use filterCollection to combine agentLabelFilter with a host or path filter for highly targeted rules.