Skip to content

[FEATURE]: Reactive security workbench for live agent execution #135

Description

@spetro511

Feature hasn't been suggested before.

I have verified this feature hasn't been suggested before.

Category

Other

Description

The Web UI should behave as a live security workbench while an agent is running, rather than limiting meaningful updates to the chat transcript. Activity, Mission, topology, memory, findings, MCP/Bolt state, and terminal surfaces should receive session-scoped updates as relevant evidence is produced.

The current Kali deployment shows the core problem: agent responses stream, but inactive workbench tabs are unmounted, several panels only poll while visible, and ordinary tool execution does not necessarily produce topology, methodology, or memory changes. The proposed design adds one session-scoped live-data channel, keeps lightweight panel state current while tabs are hidden, surfaces freshness/running indicators in tab labels, and preserves approval gates for active network actions.

Acceptance criteria:

  • Activity updates continuously during agent and tool execution, including reconnect recovery.
  • Mission, topology, memory, findings, MCP/Bolt, and terminal indicators update when their backing data changes.
  • Hidden panels do not need wasteful independent polling to remain current.
  • Opening a panel shows current state immediately with visible freshness/error status.
  • The design is validated on the Linux/Kali Web UI and does not run active scans without explicit approval.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

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