Skip to content

Conversation

@GangGreenTemperTatum
Copy link
Contributor

@GangGreenTemperTatum GangGreenTemperTatum commented Dec 10, 2025

feat: add custom guidance to summarize_when_long hook

Key Changes:

  • suggestion PR enhances summarize_when_long hook with a customizable prompt appended

  • The base prompt defines the structure and format

  • If you provide guidance, rigging injects it as additional instructions

  • The model follows both: the structure and custom guidance

    Example usage:

  summarize_when_long(
      model="groq/openai/gpt-oss-120b",
      max_tokens=80000,
      min_messages_to_keep=2,
      guidance="Focus on security testing: vulnerability findings, exploit payloads, HTTP evidence, and verifier confirmations."
  )

Changed:

  • Updates to existing hook

Generated Summary:

  • Introduced a new optional parameter guidance to the summarize_when_long function for enhanced control over the summarization process.
  • Updated the function docstring to include details about the new guidance parameter.
  • Modified the configuration to allow setting of the guidance parameter with a default value of an empty string.
  • Passed the guidance parameter during the summary generation call to the summarize_conversation function.
  • These changes enhance the flexibility and effectiveness of the conversation summarization feature.

This summary was generated with ❤️ by rigging

@monoxgas
Copy link
Contributor

Love it, nice work

@monoxgas monoxgas merged commit 015071c into main Dec 10, 2025
8 checks passed
@monoxgas monoxgas deleted the ads/cap-745-verifiers-add-guidance-param-to-summarize_when_long-hook branch December 10, 2025 19:22
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants