[FEATURE] Add Masking Function for Sensitive Data Before set_attribute

[jjayp4rk]

Problem Statement

Currently, when using set_attribute with Langfuse integration in strands-agent, there's no built-in mechanism to mask sensitive data before it's logged or traced. This creates potential security and privacy concerns when dealing with:

• API keys and credentials
• PII (Personally Identifiable Information)
• Customer data
• Internal system information
• Other sensitive attributes

Proposed Solution

Add a mask() function (or similar masking utility) that can be called before set_attribute() to automatically mask or redact sensitive information. This would provide a standardized, secure way to handle sensitive data in agent attributes.

Use Case

1. Langfuse Tracing: Mask sensitive data before it's sent to Langfuse for observability
2. Debugging: Allow developers to debug with realistic data shapes without exposing actual values
3. Compliance: Meet security and privacy requirements (GDPR, SOC2, etc.)

Alternatives Solutions

No response

Additional Context

No response

[dbschmigelski]

Hi, thanks for this and sorry for the delay.

I added a comment under #1292 which has a plan of action already detailed. @poshinchen is going to be managing the change!