Research Agent Guardrails

[sahilds1]

Title

TODO: Add "safety filtering"

OpenAI Guardrails
https://openai.github.io/openai-guardrails-python/

OpenAI Privacy Filtering
https://openai.com/index/introducing-openai-privacy-filter/

Background

Existing Behavior

Acceptance Criteria

• []

Approach

References

437 Build an input sanitizer

Risks and Rollback

Screenshots / Recordings