Skip to content

[Feature] Add deterministic, documented flows for pipe filtering or full output retrieval #58

Description

@sirknigget

I let my agent benchmark and evaluate boost with real life command outputs to see if it's suitable for my environment.

The problem right now - it's too much of a black box.
Init -> install -> let it do what it's doing internally with no control over the process and no custom hooking.

I would like to integrate good command output filters as part of my own custom hook.
Not a closed course do-it-all CLI that internally manipulates commands and doesn't provide a guaranteed mechanism for keeping raw output.

Ideal feature set:

  • 'pipe' mode, like in RTK, that lets you explicitly specify a known filter type and provide full command output, get the filtered result back.
  • documented, contract level way to get the full original output of the provided command, without hacking into the CLI and how it works.
  • safe mode which prevents Boost from doing any command modifications - ONLY routing on the exact command provided (excluding bypassing common shell mechanisms such as prefixes, "cd ..", "&&", etc.)

Overall, I currently can't use it as part of my already existing hooks. If you open up the underlying components and make them reusable, I can take advantage of it.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions