docs(examples): secure agent memory with Hindsight

[DK09876]

Adds an example showing how to use Superagent Guard + Redact to protect an agent's long-term memory — blocking prompt-injection payloads and stripping PII before content is stored, and guarding queries before recall.

It uses Hindsight (open-source memory engine) via the hindsight-superagent SafeHindsight wrapper, which applies Superagent safety checks around each memory operation. Directly analogous to the existing secure-rag-pipeline example, for the memory write/read path.

• New page: docs/content/docs/sdk/examples/secure-agent-memory.mdx

Python example (uses your Python SDK under the hood). Happy to adjust framing or move it.

---

Note

Low Risk
Documentation-only addition with no runtime, auth, or application code changes.

Overview
Adds a new SDK docs example page secure-agent-memory.mdx that walks through protecting long-term agent memory with Superagent Guard and Redact via Hindsight’s SafeHindsight wrapper (hindsight-superagent).

The page covers prerequisites, install/env setup, a full retain / recall Python sample (guard + redact on write, guard on read), a short flow diagram, retain_batch behavior (whole batch aborted if Guard blocks any item), and links to PyPI, integration source, and the Python SDK.

Note: The diff only adds the MDX file; it does not update docs/content/docs/sdk/meta.json, so the page may not appear in the examples nav until that is added separately.

^{Reviewed by Cursor Bugbot for commit 13037a6. Bugbot is set up for automated code reviews on this repo. Configure here.}

[vercel]

@DK09876 is attempting to deploy a commit to the Superagent Team on Vercel.

A member of the Team first needs to authorize it.

[DK09876]

Hi @homanp — friendly bump on this one. It adds a docs example for securing agent memory with Hindsight. Happy to make any changes you'd like; just let me know if anything needs adjusting before merge. Thanks!