Skip to content

Investigate Praxis integration for LLM request interception #17

Description

@eranra

Context

We've been engaged with the Praxis team regarding potential infrastructure collaboration. See: praxis-proxy/praxis#481

Proposal

Praxis offers a proxy solution to intercept and manipulate LLM requests and responses. Instead of maintaining separate proxy infrastructure for Skillberry-Agent's request/response enhancement capabilities, we should investigate whether we can integrate with Praxis to:

  • Intercept LLM requests and responses dynamically
  • Apply optimized assets from our skills store
  • Reduce infrastructure duplication between projects

What needs investigation

  1. Technical feasibility of integrating Praxis with our current LangGraph + MCP pipeline
  2. API compatibility and extension points
  3. Performance implications
  4. Whether Praxis can support the dynamic optimization use cases in Skillberry-Agent
  5. Timeline and effort required for integration

Related

Activity

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

Metadata

Metadata

Assignees

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