Skip to content

.Net Agent - Expose internal dialog and max-depth for plug-in model #4313

New issue
New issue
Closed
#5725
Closed
.Net Agent - Expose internal dialog and max-depth for plug-in model#4313
#5725
Assignees
crickman
Labels
.NETIssue or Pull requests regarding .NET codeagentsexperimentalAssociated with an experimental featurefollow upIssues that require a follow up from the community.sk team issueA tag to denote issues that where created by the Semantic Kernel team (i.e., not the community)

Description

@crickman
Chris (crickman)
opened on Dec 15, 2023

When agents are interacting as plugins for another agent, the internal dialog isn't exposed. While this is fine for the initial experimental phase, it warrants further refinement.

Also consider a "max-depth" concept for how long this type of conversation is allowed to run.

Activity

  1. Chris (crickman)
    added
    enhancement
    .NETIssue or Pull requests regarding .NET code
    experimentalAssociated with an experimental feature
    on Dec 15, 2023
  2. Chris (crickman)
    self-assigned this
    on Dec 15, 2023
  3. Chris (crickman)
    added this to Semantic Kernelon Dec 15, 2023
  4. Shawn Callegari (shawncal)
    added
    triage
    on Dec 15, 2023
  5. Mark Wallace (markwallace-microsoft)
    added
    sk team issueA tag to denote issues that where created by the Semantic Kernel team (i.e., not the community)
    and removed
    triage
    on Dec 15, 2023
  6. github-merge-queue
    added a commit that references this issue on Dec 15, 2023

    .Net Better repeatability for Agent KernelExample #72 (Example72_Agen…

    0ef7cf4
  7. Mads Bolaris (madsbolaris)
    added
    agents
    and removed
    enhancement
    on Jan 2, 2024
  8. Mads Bolaris (madsbolaris)
    moved this to Backlog in Semantic Kernelon Jan 10, 2024

  9. joslat commented on Jan 27, 2024

    @joslat
    Jose Luis Latorre Millas (joslat)
    on Jan 27, 2024
    Contributor

    This is much needed - thanks Chris (@crickman) for bringing this forward!
    Aside from maxdepth, how about the number of "interactions" too? that would be very convenient to identify and avoid never ending feedback loops...

  10. crickman commented on Jan 29, 2024

    @crickman
    Chris (crickman)
    on Jan 29, 2024
    ContributorAuthor

    I hear you!

  11. crickman commented on Mar 25, 2024

    @crickman
    Chris (crickman)
    on Mar 25, 2024
    ContributorAuthor

    Currently addressing in active Agent ADR: https://github.com/microsoft/semantic-kernel/blob/adr_agent_architecture/docs/decisions/0039-agent-interactions.md

  12. Chris (crickman)
    moved this from Backlog to Sprint: Planned in Semantic Kernelon Mar 25, 2024
  13. Chris (crickman)
    added
    follow upIssues that require a follow up from the community.
    on Mar 25, 2024

  14. crickman commented on Apr 2, 2024

    @crickman
    Chris (crickman)
    on Apr 2, 2024
    ContributorAuthor

    #5725

  15. Chris (crickman)
    linked a pull request that will close this issue.Net - Introducing AgentGroupChat (Step #2) #5725on Apr 2, 2024
  16. Chris (crickman)
    closed this as completedin #5725on Apr 17, 2024
  17. github-project-automation
    moved this from Sprint: Planned to Sprint: Done in Semantic Kernelon Apr 17, 2024
  18. Bryan (Bryan-Roe)
    added a commit that references this issue on Oct 6, 2024

    .Net Better repeatability for Agent KernelExample #72 (Example72_Agen…

    48a76c2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

.NETIssue or Pull requests regarding .NET codeagentsexperimentalAssociated with an experimental featurefollow upIssues that require a follow up from the community.sk team issueA tag to denote issues that where created by the Semantic Kernel team (i.e., not the community)

Type

No type

Projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions