Skip to content

Agent Memory Guard

v0.3.0 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 3d AI Agents & Assistants
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agentic-ai ai-agents ai-safety autogen crewai langchain
+10 more
llm-agents llm-security mem0 memory-poisoning openai-agents owasp prompt-injection python rag-security security

Summary

AI summary

Introduces provenance-based memory classification, multiple integration adapters, and CI fixes across feat, ci, and fix modules.

Changes in this release

Feature Medium

Adds CLI scanner, API server, ML detection, new detectors, and MkDocs documentation.

Adds CLI scanner, API server, ML detection, new detectors, and MkDocs documentation.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds Jupyter notebook examples and attack simulation demos.

Adds Jupyter notebook examples and attack simulation demos.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds GitHub Action for scanning agent memory vulnerabilities.

Adds GitHub Action for scanning agent memory vulnerabilities.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds LlamaIndex integration adapter.

Adds LlamaIndex integration adapter.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds CrewAI integration adapter.

Adds CrewAI integration adapter.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds Prometheus metrics exporter.

Adds Prometheus metrics exporter.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Feature Medium

Adds Policy.tiered() preset with default memory class taxonomy.

Adds Policy.tiered() preset with default memory class taxonomy.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Bugfix Low

Fixes CI failure from Policy.tiered() merge conflict and lint errors.

Fixes CI failure from Policy.tiered() merge conflict and lint errors.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Bugfix Low

Fixes mypy --strict errors by adding py.typed marker.

Fixes mypy --strict errors by adding py.typed marker.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Bugfix Low

Fixes ruff I001 import‑sort failures in CI.

Fixes ruff I001 import‑sort failures in CI.

Source: llm_adapter@2026-06-11

Confidence: high

 —
Full changelog

What's Changed

  • feat(classification): provenance-based memory classes and promotion rules by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/25
  • Update README.md by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/28
  • ci: fix ruff I001 import-sort failures by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/29
  • docs: clearer install CTAs, framework recipes, issue templates by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/30
  • feat(asi06): source_class, self-reinforcement detector, receipt_uri, retire_if by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/31
  • fix(ci): drop deprecated typing.Deque, fix import block order by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/32
  • chore: ignore coverage artifacts by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/33
  • fix: add py.typed marker and fix mypy --strict errors by @Metbcy in https://github.com/OWASP/www-project-agent-memory-guard/pull/11
  • feat: add Jupyter notebook examples and attack simulation demos (Issue #7) by @yarrbakr in https://github.com/OWASP/www-project-agent-memory-guard/pull/17
  • feat: add GitHub Action for scanning agent memory vulnerabilities by @hesam-oxe in https://github.com/OWASP/www-project-agent-memory-guard/pull/18
  • feat: add LlamaIndex integration adapter by @hesam-oxe in https://github.com/OWASP/www-project-agent-memory-guard/pull/19
  • feat: add CrewAI integration adapter by @hesam-oxe in https://github.com/OWASP/www-project-agent-memory-guard/pull/20
  • feat: add Prometheus metrics exporter by @hesam-oxe in https://github.com/OWASP/www-project-agent-memory-guard/pull/21
  • feat: add Policy.tiered() preset with default memory class taxonomy by @hesam-oxe in https://github.com/OWASP/www-project-agent-memory-guard/pull/23
  • fix: resolve CI failure from Policy.tiered() merge conflict and lint errors by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/34
  • v0.3.0 — CLI scanner, API server, ML detection, new detectors, MkDocs docs by @vgudur-dev in https://github.com/OWASP/www-project-agent-memory-guard/pull/38
  • docs: add Google-style docstrings to all public classes and methods by @soham31415 in https://github.com/OWASP/www-project-agent-memory-guard/pull/39

New Contributors

  • @vgudur-dev made their first contribution in https://github.com/OWASP/www-project-agent-memory-guard/pull/25
  • @Metbcy made their first contribution in https://github.com/OWASP/www-project-agent-memory-guard/pull/11
  • @yarrbakr made their first contribution in https://github.com/OWASP/www-project-agent-memory-guard/pull/17
  • @hesam-oxe made their first contribution in https://github.com/OWASP/www-project-agent-memory-guard/pull/18
  • @soham31415 made their first contribution in https://github.com/OWASP/www-project-agent-memory-guard/pull/39

Full Changelog: https://github.com/OWASP/www-project-agent-memory-guard/compare/v0.2.2...v0.3.0

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track Agent Memory Guard

Get notified when new releases ship.

Sign up free

About Agent Memory Guard

All releases →

Related context

Beta — feedback welcome: [email protected]